British Journal of Medical Psychology (2000), 73, 243–257 Printed in Great Britain 243

q 2000 The British Psychological Society

The personality and cognitive-epistemological

traits of cognitive-behavioural and psychoanalytic
psychotherapists
A ndrew R. A rthur*
Chartered Clinical Psychologist, Psychotherapist

This is an investigation into the personality and cognitive-epistemological traits of

psychotherapists from two major psychotherapeutic orientations. The purpose is to
examine whether there are distinctive trait patterns associated with each orientation.
Two hundred and forty-seven psychotherapists from the psychoanalytic and cognitive-
behavioural orientations completed standardized personality and epistemological trait
inventories.The results reveal signiŽcant differences on these measures between the two
orientation groups, and suggest that different patterns of personality and cognitive-
epistemological traits are associated with practitioners from these two orientations.
These trait Žndings are summarized into two comparative descriptions, which illustrate
their differences.

Theoretical orientation informs the type of treatment a patient receives; yet relatively
little is known concerning the reasons psychotherapists practise within different
orientations. Some authors (Cummings & Lucchese, 1978; Lazarus, 1978; Schwartz,
1978) suggest that the major inuences are environmental, for example training,
supervision, economic, and clinical experience but most believe the therapist’s person-
ality, and cognitive-epistemological traits (styles of thinking and theories knowledge)
are the major determining factors (e.g. Scandell, Wlazelek, & Scandell, 1997; Tremblay,
Herron, & Schultz, 1986; Vasco, Garcia-Marques, & Dryden, 1993).
There is little published literature concerning personality, cognitive-epistemology and
orientation, only 43 articles were located on the subject. The majority (31) were
speculative, theoretical or anecdotal. Nineteen were found in a special edition of
Psychotherapy: Theory, Research and Practice where the editor concludes, ‘Without an
understanding of theory we cannot practice knowledgeably. And without an under-
standing of personality we cannot understand the source and development of theory’
(Barron, 1978, p. 307). Eleven were based upon empirical research and Žnd evidence of an
association between orientation, personality and cognitive-epistemological traits (Angelos,
1977; Caine & Smail, 1969; Hill & O’Grady, 1985; Johnson, Germer, Efran, & Overton,
1988; Keinan, Almagor, & Ben Porath, 1989; Kolevzon, Sowers-Hoag, & Hoffman, 1989;
Kreitman, 1962; Schacht & Black, 1985; Tremblay et al., 1986; D. E. Walton, 1978; H. J.
Walton, 1966). As part of the present investigation, and for the Žrst time, the 11 empirical
*Requests for reprints should be addressed to Dr Andrew R. Arthur, Chartered Clinical Psychologist, Psychotherapist,
1 Langton Close, Battle, East Sussex TN33 0XH, UK.
244 Andrew R. Arthur
studies were analysed for the signiŽcant trait data they contained about psychotherapists
from two major orientations, psychoanalytic and behavioural. The behavioural group did
not include cognitive-behaviourists because the studies were performed before this group
had developed its distinctive identity. From this analysis all of the statistically signiŽcant
personality and cognitive-epistemological traits found were collated into descriptions of
these two orientations.
The studies found that behaviourists tended to be characterized by the following
traits: they are not predominantly thinkers about the internal/inner world or its
complexity; to investigate their hypotheses they require physical-sensory data that are
concrete, objective, observable and measurable; they consider environmental factors as
very important causes of behaviour; behaviourists see themselves as rational and
empirical; as therapists they prefer to set limits, look for change and want to reinforce
it; they are more concerned about thoughts than feelings; a lower acceptance of aggressive
client feeling, intimate contact and awareness of feeling reactivity is reported, as is a
liking for stability, realism and breaking down phenomena into elements; they can be
down to earth, conventional, inartistic, traditional, predictable, orderly, stable and
realistic; they rate themselves as active, having initiative, and with being practical,
assertive, dominant and extrovert.
The psychoanalytic psychotherapist’s thinking is primarily concerned with the inner
world: there is a tendency to depression, moodiness, anxiety and introversion; the main
thinking approach is intuitive, having ideas, imagination, theorizing and experimenting;
in therapy, psychoanalytic psychotherapists are concerned with the intrapsychic, dreams,
memories, and free association; they see themselves as serious, complex, and having
metaphysical thoughts; as therapists they are concerned with feelings and insight not
with focus and change, they have a feeling reactivity, acceptance of aggression and
capacity for contact; a uid, changing, creative, non-conforming, imaginative, indivi-
dualistic and active personality style is seen; change is encouraged and there is a tendency
to seeing the whole picture rather than its parts; they see themselves as passive,
impractical, non-assertive and reactive rather than pro-active.
Methodological problems found in the studies suggested that any satisfactory study
of personality and cognitive-epistemological trait differences between psychotherapists
of different orientations should employ: (a) at least 50 participants in each orientation
sub-group; (b) participants chosen from recognized, high quality and orientation-
committed professional bodies; (c) an internal check on psychotherapeutic orientation
and commitment; (d) standardized, valid and reliable measures of personality and
epistemology; and (e) an analysis of the effects of gender and experience.
These criteria formed the basis for the unique design and improved methodology of
the present investigation. The aim was to investigate personality and cognitive-
epistemological style differences between psychoanalytic and cognitive-behavioural
psychotherapists using well-researched and established measures with proven reliability
and validity. The participants were from Žrmly established professional bodies, with
well-developed training, and conŽrmed their orientation by answering verifying com-
mitment questions. There was a large enough sample to be able to analyse subgroups
statistically in order to explore the interactions of orientation with gender and experience.
The important effects of these variables on scores had not been previously evaluated in
studies.
 Personality of psychoanalytic psychotherapists 245
The measures

The literature review found that a number of personality inventories had been used to
measure differences between psychotherapists of different orientations (e.g. Cattell 16PF,
California Psychological Inventory, and Meyers Briggs Type Indicator). A relatively
new personality inventory, The Millon Index of Personality Styles (MIPS; Millon, 1994),
was chosen for this study because its 24-trait structure provided the comprehensive and in-
depth assessment of personality and cognitive styles required to detect and describe subtle
differences between psychotherapists. Additionally, as the MIPS is pan-theoretical, and
therefore its items are derived from major personality theories and theorists (Freud, Jung,
Sullivan, Murray, Leary, Bateson) and DSM–IV (American Psychiatric Association, 1994), it
was thought more likely to reect the diversity of participants’ personality and cognitive
styles than a factor-analytically derived instrument like the NEO-5 (Costa & McCrae, 1992).
Two instruments were chosen to measure epistemic belief, the Organicism-Mechanism
Paradigm Inventory (OMPI; Johnson, Howey, Reedy, Gribble, & Ortiz, 1988) and the
Psycho-Epistemological ProŽle (PEP; Royce & Mos, 1980). The OMPI was selected from
a major review of seven measures of epistemological style conducted by Johnson
and Miller (1990). They performed a factor analysis on these instruments and found
evidence to support their hypothesis that there exist two dichotomous epistemologies
‘. . . associated with a distinctive cognitive and personality style. Mechanists, who endorse
an objectivist worldview, appear to be conscientious, but somewhat rigid, restricted, and
anti-intellectual. Organicists, who endorse a subjectivist worldview, are more humani-
tarian, interpersonally competent, and intellectually open’ (Johnson & Miller, 1990, p.1).
The OMPI (Johnson et al., 1988) is designed to measure an individual’s preference for one
of these two epistemological styles; Organicism or Mechanism. When the OMPI was
used to study the epistemological styles of groups consisting of administrators, educators,
academic psychologists and psychotherapists it was found that the ‘ . . . predicted
relationships between the OMPI and theoretical orientation, educational philosophy,
vocational interests and counselling philosophy were empirically conŽrmed’ (Johnson
et al., 1989, p.2).
The second epistemological instrument, the PEP (Royce & Mos, 1980), was selected
because it measures epistemic belief from a different philosophical perspective than the
OMPI. It is based upon the philosophical system of J. R. Royce (1964) who postulated,
and showed evidence for the existence of three basic theories of knowledge; Metaphorism,
Rationalism and Empiricism. Royce and Mos (1980) demonstrated that when the PEP
was given to groups involving 947 participants from different professional/orientation
groups, the pattern of scores did proceed in the predicted direction for predominant
epistemic belief. Different epistemological styles were found to be associated with
particular professional groupings and orientations.
For the purpose of this study epistemic belief is differentiated from cognition, which is
measured by the cognitive modes scales of the MIPS. Cognition is deŽned as the structure
and functional style of the cognitive apparatus itself and epistemology as the beliefs,
theories and philosophies of knowledge that arise from that apparatus. Because the
cognitive and epistemic measures employed in this study do appear to represent the two
cognitive-epistemological styles of the therapy orientations noted in the 11 studies
reviewed earlier, the fundamental assessment of these traits is limited.
246 Andrew R. Arthur
Hypothesis
Trait data from the 11 reviewed studies and the research-therapeutic practices of the two
orientations suggested signiŽcant differences would be found on personality and
cognitive-epistemological inventories (MIPS, PEP and OMPI). It was hypothesized
that the cognitive-behaviourist would score signiŽcantly higher on traits associated with:
(a) a rational, empirical, objective and reductionistic cognitive-epistemology; and (b) an
optimistic, conŽdent, assertive, extravert, orderly, and stable personality style. The
psychoanalytic psychotherapist, it was hypothesised, would score signiŽcantly higher on
traits associated with: (a) an intuitive, imaginative, symbolic, and holistic cognitive-
epistemology; and (b) an introverted, passive, unassertive, reactive and unstable
(depressive and anxious) personality style.

Method

Participants
All cognitive-behavioural (CBT) participants were British Psychological Society Registered Chartered
Clinical Psychologists who were either members, or accredited members (and therefore United Kingdom
Council of Psychotherapists (UKCP) registered) of the British Association of Behavioural and Cognitive
Psychotherapy (BABCP). All accredited BABCP members (93) were invited to participate. One hundred and
forty-four clinical psychologists who were BABCP members were randomly selected from the register of 1004.
An additional opportunist sample of 34 potential participants was provided by two senior Chartered Clinical
Psychologists who were well known for their commitment to CBT.
All psychoanalytic psychotherapists were registered with the psychoanalytic section of the UKCP and/or
with the British Confederation of Psychotherapists(BCP)—who maintain a register of only psychoanalystsand
psychoanalytic psychotherapists. The invited psychoanalytic participants consisted of an opportunist sample
(N = 104) and a random sample (N = 169). The opportunistsample consistedof psychotherapistswho had some
professionalcontact with the author. The random sample was selected from the professionalregister of the BCP.
The different methods of sampling produced different participation rates. Random sampling produced a
participation rate of 38% (90 of 237 invited) for cognitive-behaviourists and 40% (68 of 170 invited) for
psychoanalytic psychotherapists. Opportunist sampling produced a higher participation rate of 68% (23 of
34 invited) for cognitive-behaviourists and 64% (66 of 103 invited) for psychoanalytic psychotherapists.
A total combined random and opportunist sample of 544 psychoanalytic and cognitive-behavioural
psychotherapists were invited by letter to participate in this study. Altogether, half of the invited sample
consented (267, 49%) and were mailed personality and cognitive-epistemological inventories to complete.
Almost all (247, 93%) returned completed and useable forms within the time-frame of the study. They were
divided almost equally between psychoanalytic(134, 54%) and cognitive-behavioural psychotherapists(113,
46%). The majority were female (146, 59%), the mean age was 48.7 years (range= 27–80) and the mean
number of years in clinical practice was 15.7 years (range= 1–59).
Participants received information that the research was a study of psychotherapists’personalities and were
asked to complete, by post, an information questionnaire, a personality and epistemological inventory, and
answer the three questions relating to their therapeutic orientation. Those participants who returned
completed and useable questionnaires were then asked to complete a further epistemological inventory (the
PEP). A reminder letter was sent to those who did not return questionnaires.

Instruments
Personality. The MIPS is a 180-item true/false questionnaire designed to measure personality and cognitive
traits in normally functioning adults. It consists of 12 bi-polar scales (24 traits) grouped into three main
personality domains: motivating aims, cognitive modes, interpersonalbehaviours,an internal reliability scale
 Personality of psychoanalytic psychotherapists 247
(Consistency), two internal validity scales (Positive and Negative Impression), and a measure of psychological
stability (Adjustment) that is derived from six personality trait scales. Although the MIPS is a recent
personality inventory it appears to have undergone a comprehensiveprogramme of theoretical, item and scale
development, followed by a normative standardization of 1000 adults. As part of the inventory’s
development Millon (1994) found that the internal consistency reliability co-efŽcient alpha median was
.78, and split-half reliability median was .82. Retest reliability was assessed on a sample of 50 adults, the
median interval was 2 months and the range was 20–82 days. The median retest reliability was found to be
r = .85. The factor structure of the MIPS was analysed at the item level by another investigator (Weiss, 1997)
who reported the presence of ‘ . . . Žve factors that were consistent with the Žve-factor model of personality’
(p.513). External validity was established between the MIPS scales and seven other major personality
inventories (e.g. Cattell 16PF, California Psychological Inventory, and NEO-5). Intercorrelational data
reported by Millon (1994) led him to conclude that ‘The patterns of convergent and divergent correlation
between the MIPS and other tests of personality reported in the section, obtained to demonstrate external
validity, are largely consonant with expectation,based upon the author’s theory of normal personology and on
the item content of the respective scales’ (p.87).

Epistemology. The OMPI (Johnson et al., 1988) is designed to measure an individual’s preference for one of two
epistemological styles: Organicism or Mechanism. It requires responses to 26 forced-choice questions, for
example: (a) The world is like a large, living organism or (b) The world is like a large, complex machine.
Johnson et al. (1988) reported the OMPI as having good internal consistency ‘ . . . with a Guttman split half
co-efŽcient of .86 and a Cronbach alpha coefŽcient of .76. A 3-week retest showed a stability coefŽcient of
.77’ (p.825). However, from Johnson et al.’s (1990) factor analysis it was noted that ‘Reliabilities for separate
Organicism and Mechanism scales scored from the OMPI were good (.85 and .78), but the factor analysis
indicated 16 strong markers of Organicism (revised reliability= .83), but only 6 strong markers for
Mechanism (revised reliability= .52)’ (p.6).
The second epistemological instrument, the PEP (Royce & Mos, 1980), was offered to participants who
had returned the Žrst set of questionnaires,67% (165 of 247) completed the inventory. The PEP is a 5-point
Likert scaled questionnaire that asks for responses to 90 questions, for example: I feel most at home in a
culture in which realism and objectivity are highly valued. The PEP in its Žnal form was standardized on
1342 participants.When the PEP was examined for construct validity, and a factor analysis performed, Royce
and Mos (1980) claimed to have identiŽed three factors that corresponded to the three epistemological styles.
However, results showed salient item loading (6 .25 or greater) for only 40% of the items, suggesting that
the remaining items needed revision for test inclusion, or the test could be shortened to 41 items. Personal
telephonic communication with Mos (May, 1997) conŽrmed this possibility. In their reported factor analysis
of seven worldview inventories (discussed earlier) Johnson and Miller (1990) also found that a fairly radical
revision of the scoring was necessary for the PEP. They found three new factors named Insight, Culture, and
Science. The reported reliabilities of these new scales were .83, .85, and .73 respectively (Johnson & Miller,
1990). Royce and Mos (1980) reported split-half reliability coefŽcients for the original PEP scales as:
Rationalism= .77, Metaphorism= .88, and Empiricism = .77. Test–retest reliabilities were reported at 3
and 9 months on small samples (19 and 43 participants). This revealed a range across scales and conditions of
.61 to .87.

Attitude to orientation. In addition, participants’ attitudes to their orientation were explored by asking three
ad hoc questions to measure commitment, satisfaction and therapeuticpractice. The three questions were: (a) I
am completely committed to my prime theoretical orientation; (b) I am completely satisŽed with my prime
theoretical orientation; and (c) In my clinical practice I remain completely within my prime theoretical
orientation. The 5-point response scale varied from 1 = not at all, to 5 = completely.

Results

Demographic variables
t-Tests were performed to investigate between-orientation differences for age and
248 Andrew R. Arthur
experience, and a chi-square analysis for differences in gender distribution. The only
statistically signiŽcant result was that psychoanalytic psychotherapists tended to be older
than cognitive-behaviourists (M = 51.9 and M = 44.9 years respectively; t(245) = 5.74,
p < .001). For the purposes of analysis, participants were divided into three approximately
equal groups according to years of experience: novice (1–10 years; N = 82), intermediate
(11–18 years; N = 81) and senior (19–59 years; N = 78).

Personality and cognitive measures

The ANOVA results (see Table 1) showed signiŽcant differences on MIPS scales for each
independent variable (orientation, gender, and experience), and one interaction (orienta-
tion ´ experience). There were 11 (10 traits and Adjustment) MIPS scale score differences
between psychoanalytic and cognitive-behavioural psychotherapists, nine (eight traits
and adjustment) between males and females, and two (one trait and Positive Impression)
between participants with different levels of experience. For the one signiŽcant factor
interaction of orientation ´ experience, there were Žve (four traits and Positive Impression)
score differences.
Table 1 presents the overall scores for each orientation. Psychoanalytic psychotherapists
scored higher than cognitive-behaviourists on Preserving, Intuiting, Feeling, and Innovat-
ing. Cognitive-behaviourists scored higher than psychoanalytic psychotherapists on Enhan-
cing, Individuating, Sensing, Thinking, Retiring, Conforming, and Adjustment.
When the effect of orientation ´ experience was analysed, 2 of the 10 MIPS trait scores
that differed by orientation (Sensing and Conforming) showed an interaction with
experience; two additional trait differences (Systematizing and Yielding) occurred only as
an interaction between orientation and experience (see Table 1). For Sensing and
Conforming, analysis of variance showed that the simple effect was between novice
therapists for Sensing (F(1,80) = 21.31, p = .000) and Conforming (F(1,80) = 21.84,
p = .000). Novice psychoanalytic psychotherapists scored lower for Sensing (M = 9.84,
SD = 5.43) and Conforming (M = 23.47, SD = 6.62) than novice cognitive-behaviour-
ists (M = 15.45, SD = 5.37) and (M = 31.58, SD = 9.09) respectively. A follow-up one-
way ANOVA also found a simple effect within the psychoanalytic group for Conforming
(F(2,127) = 5.63, p = .005). A post hoc Tukey HSD comparison (M difference = 5.71,
p = .003) showed that novices scored lower (M = 23.47, SD = 6.62) than seniors
(M = 29.18, SD = 9.29). Two trait differences (Systematizing and Yielding) occurred
only as an interaction between orientation and experience. Analysis of variance for these
two traits showed that the simple effect was between novice therapists for Systematizing
(F(1,80) = 7.60, p = .007), and Yielding (F(1,80) = 4.07, p = .047). Novice psycho-
analytic psychotherapists scored lower for Systematizing (M = 28.22, SD = 8.66) than
novice cognitive-behaviourists (M = 34.06, SD = 10.42); and novice psychoanalytic
psychotherapists scored higher for Yielding (M = 18.88, SD = 6.66) than novice cognitive-
behaviourists (M = 15.85, SD = 6.68). A follow-up one-way ANOVA also found a simple
effect within the psychoanalytic group for Yielding (F(2,127) = 3.91, p= .02). A post hoc
Tukey HSD comparison (M difference = 3.85, p = .02) showed that novices scored higher
(M = 18.88, SD = 6.66) than seniors (M = 15.03, SD 6.04). Findings for the interaction
between orientation and experience for the MIPS suggests that novice psychoanalytic
psychotherapists are less Systematizing, Sensing, Conforming and more Yielding,
Table 1. Means and standard deviations for MIPS scores, and analysis of variance by orientation, gender and experience
Psychoanalytic Cognitive
psychotherapists behaviourists
N = 130 N = 111 Orientation
Orientation Gender Experience ´ experience
M SD M SD F(1,239) F(1,239) F(2,238) F(2,238)
Enhancing 23.68 7 25.80 7.28 6.619* 2.983 0.552 1.223
Preserving 14.59 7.92 11.74 8.34 7.83** 1.849 0.222 0.719
Modifying 24.17 8.57 26.05 9.32 1.821 4.222* 0.543 1.121
Accommodating 17.08 8.87 16.52 9.39 0.128 4.963* 0.554 0.388
Individuating 13.71 6.53 15.42 6.28 4.262* 3.107 0.021 0.517
Nurturing 28 7.03 27.13 7.90 0.735 3.452 0.98 0.737
Extraversing 26.51 7.92 26.61 9.00 0.006 10.999** 0.198 2.496
Introversing 9.58 6.37 9.57 6.80 0 4.594* 0.313 1.927
Sensing 11.08 5.63 14.06 6.36 14.411*** 0.27 0.137 3.382*
Intuiting 24.03 7.49 19.95 8.94 12.729*** 0.015 0.222 2.515
Thinking 10.26 5.72 15.68 6.71 45.127*** 12.848*** 3.788* 2.661
Feeling 28.56 6.87 24.38 8.34 16.509*** 3.312 0.998 0.627
Systematizing 30.51 9.45 32.32 9.83 1.703 2.862 0.763 3.192*
Innovating 29.54 9.22 26.56 9.64 5.095* 1.147 0.692 2.481
Retiring 13.73 7.34 16.41 10.02 6.695** 18.057*** 0.729 0.229
Outgoing 29.33 9.89 29.82 11.74 0.009 0.102 0.844 0.31
Hesitating 16.5 9.03 14.34 9.02 2.499 0.603 1.547 1.667
Asserting 26.17 8.13 28.48 9.82 2.994 0.915 2.064 2.967
Dissenting 17.57 6.28 17.13 7.11 0.161 3.956* 0.215 1.566
Conforming 25.75 8.23 29.89 8.49 13.038*** 0.004 0.455 5.98*
Personality of psychoanalytic psychotherapists

Yielding 17.04 6.58 15.91 6.40 1.687 1.806 1.616 3.771**

Controlling 15.42 5.61 16.00 6.70 0.287 0.411 0.866 0.272
Complaining 16.82 7.2 15.90 7.90 1.127 11.323*** 0.959 0.192
Agreeing 34.35 8.07 33.72 8.69 0.23 0.333 0.677 0.784
Positive impress 1.15 1.38 1.50 1.41 3.68 0.67 3.218* 3.742*
Negative impress 1.65 1.45 1.48 1.52 0.673 0.932 0.21 1.295
Consistency 3.75 0.99 3.96 1.09 2.295 0.203 0.325 0.621
Adjustment 46.57 7.86 48.95 8.32 5.784* 4.53* 0.384 1.583
*p< .05; **p< .01; ***p< .001.
Note: There were no signiŽcant between participant effects found on MIPS scale scores for: orientation ´ gender, gender ´ experience, and orientation ´ gender ´ experience.
249
250 Andrew R. Arthur
compared with their novice cognitive-behaviourist counterparts, and that within their
own orientation, senior psychoanalytic colleagues are more Yielding and less Conforming
than novices.
To summarize, when trait scores are examined within their MIPS personality domains
for orientation and orientation ´ experience, the following pattern is found:

(1) 75% (6/8) cognitive modes traits differ by orientation;

(2) 50% (3/6) motivating aims traits differ by orientation;
(3) 30% (3/10) interpersonal behaviour traits differ by orientation.

These Žndings suggest that cognitive modes and motivating aims are more important
than interpersonal behaviour in discriminating therapeutic orientation membership.

Epistemological measures
When the OMPI was evaluated ANOVA results showed signiŽcant differences for
orientation (F(1,238) = 7.41, p = .007) and orientation ´ experience (F(2,237) = 6.21,
p = .002). There was a signiŽcant difference between psychoanalytic psychotherapists and
cognitive-behaviourists on the OMPI score. Psychoanalytic psychotherapists scored
higher (M = 20.98, SD = 2.84) and therefore towards the Organicism direction. Cogni-
tive-behaviourists scored lower (M = 19.84, SD = 3.64) and therefore towards the
Mechanism direction.
When the effect of orientation ´ experience was analysed, a one-way ANOVA revealed
a simple effect between novices from both orientations (F(1,81) = 23.50, p = .000).
Novice psychoanalytic psychotherapists scored higher (M = 21.84, SD = 2.42) than
novice cognitive-behaviourists (M = 18.67, SD = 3.55). A follow-up one-way ANOVA
also found a simple effect within the psychoanalytic group (F(2,127) = 3.92, p = .02). A
post hoc Tukey HSD comparison (M difference = 1.52, p = .03) showed that novices scored
higher (M = 21.84, SD = 2.42) on the OMPI than seniors (M = 20.32, SD = 3.46).
These Žndings suggest that novice psychoanalytic psychotherapists are more Organismic
than novice cognitive-behaviourists, and their own senior psychoanalytic colleagues.
The ANOVA results for the second epistomological instrument (PEP), completed by
two-thirds (165) of participants, showed signiŽcant differences for orientation (Table 2).
Results showed psychoanalytic psychotherapists scored signiŽcantly higher on Culture
and Metaphorism than cognitive-behaviourists (Table 2).
Analysis of variance showed an interaction between orientation and experience for one
scale, Insight (F(2,162) = 4.60, p = .01). A one-way follow-up ANOVA revealed that a
difference between senior psychoanalytic psychotherapists and cognitive-behaviourists
just failed to reach signiŽcance (F(1,52) = 3.91, p = .053). Senior psychoanalytic
psychotherapists scored higher (M = 69.88, SD = 6.45) for Insight than senior cognitive-
behaviourists (M = 66.86, SD = 4.85). A follow-up one-way ANOVA found a simple
effect within the psychoanalytic group (F(2,87) = 4.23, p = .02). A post hoc Tukey HSD
comparison (M difference = 4.58, p = .02) showed that seniors scored higher for Insight
(M = 69.88, SD = 6.45) than novices (M = 65.31, SD = 5.8). These Žndings suggest
that very experienced psychoanalytic psychotherapists score higher for Insight than less
experienced colleagues.
Table 2. Means and standard deviations for PEP scores, and analysis of variance by orientation, gender and experience
Psychoanalytic Cognitive
psychotherapists behaviourists Orientation
N = 90 N = 75 Orientation ´ experience
Orientation Experience ´ experience ´ gender
M SD M SD F(1,163) F(2,162) F(2,162) F(2,162)
Culture 60.21 7.58 55.83 8.85 9.547** 0.108 0.776 4.525*
Empiricism 93.01 10.89 95.67 9.31 1.972 3.296* 1.907 1.821
Insight 66.82 6.65 67.99 5.91 0.872 0.952 4.603* 0.749
Metaphorism 108.52 11.01 102.2 11.95 10.629*** 0.706 0.24 0.918
Rationalism 95.9 11.13 98.31 10.21 2.06 2.658 2.739 1.712
Science 24.83 5.36 26.33 5.49 2.012 3.077* 0.45 2.306
*p< .05; **p< .01; ***p < .001.
Note: There were no signiŽcant between participant effects found on PEP scores for: gender, orientation ´ gender, or experience ´ gender.
Personality of psychoanalytic psychotherapists
251
252 Andrew R. Arthur
Orientation commitment questions. The mean values and standard deviations of responses to
the three 5-point scale attitude to orientation commitment questions are presented in
Table 3. ANOVAs showed a main effect for the variable orientation for all three questions
separately and the total orientation score, the sum of all three questions,
(F(1,244) = 17.38, p = .000).

Table 3. Commitment question scores for the two orientations

Orientation
Analyst Behaviourist
N = 133 N = 112
M SD M SD
Attitude
Commitment 3.91 0.94 3.49 1.09
Practice 3.47 1.11 2.71 1.12
Satisfaction 3.68 0.97 3.38 0.99
Total score 11.07 2.72 9.57 2.76

Psychoanalytic psychotherapists responded with a higher score (in more agreement)

than cognitive-behaviourists to the three questions concerning their commitment,
practice and satisfaction within their orientation. This suggests psychoanalytic psy-
chotherapists remain more within their orientation, are more committed to it and
satisŽed than cognitive-behaviourists. The three questions were also used as a validity
check on whether participants were committed, satisŽed and practised within their
professed orientation. The range of scores possible on the total of all three questions was
3–15. The mean total score was 10.37 (SD = 2.84) for both orientations, suggesting that
the majority of participants were probably committed, satisŽed, and practising within
their professed theoretical model.
It was also important to conŽrm that participants perceived themselves as belonging to
their registered orientation. Over 96% selected from a list the same orientation as their
professional memberships indicated, strongly suggesting that participants were correctly
identiŽed by orientation. However, the absence of psychometric data for these three
questions limits the conclusions that can be drawn.

Discussion
Overall, the results appear to support the hypothesis that committed and trained
psychotherapists from psychoanalytic and cognitive-behavioural orientations would
show differences on personality and cognitive-epistemological measures congruent
with traits identiŽed from the previous research review. A number of these differences
was found to exist between inexperienced participants from the two orientations.
Although there are clear differences between the two orientations, it is also of note that
they share some similarities (see Table 1). These occur mainly within the interpersonal
behaviours domain of the personality inventory (MIPS) and suggest that psychotherapists
 Personality of psychoanalytic psychotherapists 253
from these two orientations share similar patterns of social behaviour. Analysis of the
MIPS cognitive domain scores suggests that both groups are also alike in the amount of
attention they pay to internal and external cues for information. Within the MIPS
motivational domain, scores show that they share a similar level of need to nurture others,
and are equally active and passive. Analysis of the epistemological measure (PEP) Žnds
that both orientations value, to a similar extent, clear thinking and analysis, and are
committed to testing the reliability and validity of observations.
On measures of personality and cognitive style, orientation accounted for 13
signiŽcant MIPS scale differences (12 traits and Adjustment). Of 24 MIPS trait scales,
12 differed signiŽcantly between psychoanalytic and cognitive-behavioural psychothera-
pists. Four of these 12 trait score differences occurred between novices from both
orientations.
An examination of the epistemological style measures also found an effect for
orientation. The OMPI showed a signiŽcant difference between psychoanalytic and
cognitive-behavioural psychotherapists in the predicted direction. Psychoanalytic psy-
chotherapists scored signiŽcantly towards Organicism and cognitive-behaviourists
towards Mechanism. However, a detailed breakdown of OMPI scores by experience
showed the difference occurred at the novice level both within and between orientations.
With more experience the two orientations then seem to share similar philosophical
worldviews. The PEP results, using Royce and Mos’ (1980) three original factors
(Rationalism, Empiricism, Metaphorism) showed cognitive-behavioural and psycho-
analytic psychotherapists to differ signiŽcantly on Metaphorism in the expected direc-
tion. However, on the other two scales differences between the groups were not observed.
Evaluating the PEP for Johnson and Miller’s (1990) three revised factors (Culture,
Science, Insight) the results showed that Culture, and nearly Insight, differed signiŽ-
cantly by orientation. There is some evidence from this investigation to support the
hypothesis that the two orientations have different epistemological styles.
The three questions purporting to measure commitment and satisfaction to orientation
showed signiŽcant differences between orientations. Psychoanalytic psychotherapists
scored higher on each question suggesting that they are more committed and satisŽed
with their orientation. This Žnding raises the possibility that the different levels of
practitioners’ commitment and satisfaction may have affected the differences obtained on
the personality and cognitive-epistemological measures. The three questions also
validated the participants’ orientation commitment; both orientations appear to
belong to their publicly professed orientation. However, self-report accounts of therapist
orientation, commitment, and practice cannot be relied upon necessarily to reect how
therapists may actually practice.
Results indicate that the orientation of the psychotherapist consistently and perva-
sively affected scores on all the measures. This strongly suggests that psychotherapists’
theoretical orientation reects certain characteristic personality traits and cognitive-
epistemological styles. Personality and cognitive-epistemological factors were found to
clearly distinguish psychoanalytic from cognitive-behavioural psychotherapists. They
have distinctly different motivational aims (MIPS), cognitive styles (MIPS), epistemo-
logical beliefs (OMPI and PEP), and, to a lesser extent, interpersonal behaviours (MIPS).
Level of commitment, and satisfaction with orientation, as measured by questionnaire
responses, were similarly affected.
254 Andrew R. Arthur
The personality and cognitive-epistemological trait descriptions and deŽnitions used
by the authors’ of the inventories are used to describe and compare the similarities and
differences found between psychoanalytic psychotherapists and cognitive-behaviourists
on the measures. The descriptions, shown below, are consistent with those derived from
the meta-analysis of personality and cognitive-epistemological trait differences (Arthur,
1998) summarized at the beginning of this article.

Orientation characteristics

Psychoanalytic psychotherapists. Compared with cognitive-behaviourists, the data obtained

suggest that this orientation tend to be characterized by the desire to avoid pain,
unpleasure or negative effect. They are especially sensitive to threats and avoid
unnecessary risk. They may be more aware of feelings of anxiety or depression, they
did not appear to be unduly active or passive, and there seems to be a balance of the drive
towards self or others.
Their cognitive-epistemological style is similar to cognitive-behaviourists in the
relative amount of attention paid to the internal and external world for information.
However, psychoanalytic psychotherapists rely predominantly on their intuition rather
than the physical senses for gathering this information. They then process their knowledge
through the use of feeling by introspective analysis, insight, and empathy. This group
often organize their knowledge in an innovative way, being prepared to step outside
cognitive structures and allow their imagination to suggest a unique understanding.
They are likely to believe that the universe is a constantly changing system and that
it is more important to understand the functioning whole than examine its parts.
By rationally examining the relationship between a person and the world through
conceptualization and thought, they believe that understanding will develop. Unlike
cognitive-behaviourists, there is a marked tendency to use symbolic and metaphorical
thought processes, and to test the validity of their perception in terms of the universality
of insight and awareness. They are more likely to remain within their theoretical analytic
model in both belief and practice. Whilst they may innovate within it, they are unlikely
to use other psychological models.
Psychoanalytic psychotherapists are not unlike the cognitive-behaviourists in actual
interpersonal behaviour. There are, however, a few differences that would distinguish
them from their behavioural colleagues. At the beginning of their careers they may
appear less conventional, orderly, responsible, proper and conscientious, but this changes
with time until they are similar in these traits to cognitive-behaviourists. Novice
psychoanalytic psychotherapists also have a tendency to servility, and adherence to the
expectations of those they follow. However, like their conformity, this difference
disappears with greater experience.

Cognitive-behaviourists. Compared with psychoanalytic psychotherapists, this group tend

to be characterized by the desire to enhance life, not just avoid unpleasure and pain. Their
primary motivating drive is to seek out opportunities to develop, foster and enrich life.
As a consequence cognitive-behaviourists may have a tendency to experience less anxiety
and depression and show more signs of psychological adjustment, emotional stability and
 Personality of psychoanalytic psychotherapists 255
cope better with stress. They are neither predominantly active nor passive in enacting this
primary drive. There is a tendency to be self-focused, self-actualized, have a strong sense
of identity and be more independent of others.
Their cognitive-epistemological style is similar to psychoanalytic psychotherapists in
the relative amount of attention paid to the internal and external world for information.
However, there is a strong sense of reliance on the physical senses for tangible, structural
and well-deŽned information gathering. This orientation prefers quantitative vs.
qualitative information, and is pragmatic and realistic. They prefer thinking to feeling,
emphasizing the use of reason, logic and a reduction of emotional input; whereas
psychoanalytic psychotherapists may rely on their feelings to understand a patient, the
behaviourist prefers observation and measurement of behaviour. They have a tendency to
try to Žt new information into their pre-existing cognitive system. This is particularly so
for less experienced clinicians. Cognitive-behaviourists are likely to be conventional,
conforming, objective and realistic in their cognitive style. Compared to psychoanalytic
psychotherapists, their epistemology is that the universe is composed of distinct parts and
knowledge must reduce these to basic units. The universe is static, and can be understood
by an external knower and therefore knowledge can be distinguished from the subject.
The scientiŽc methodology is empirical, and emphasis is placed on sense data over
concepts. A tendecy to stray from the cognitive-behavioural model is suggested by their
responses to the ad-hoc attitude to orientation questions.
Cognitive-behaviourists have less need to give or receive affection, have relationships
or express feelings in their interpersonal relationships than psychoanalytic psychothera-
pists. Less experienced cognitive-behaviourists are more conventional, orderly, proper and
conscientious. However, they are no more controlling in their interpersonal relationships
than psychoanalytic psychotherapists. This may not, of course, necessarily reect their
clinical behaviour.

Limitations of the study and suggestions for further research

The conclusions that can be drawn from this investigation are limited by a number of
factors, including the sample, the instruments, and the methodology.
The method of sampling limits conclusions in three ways. First, psychotherapists were
not invited to participate in this study by completely random means. Opportunist
samples were used to increase participation rates, and it was observed that they were more
likely to participate in the study than randomly selected participants. A professional
connection with the author may have affected the opportunist sample’s responses,
although steps were taken to minimize this. Second, only two (albeit major) orientations
were compared in this investigation. Therefore, although personality and cognitive-
epistemological trait differences were measured, the Žndings may not generalize to other
psychotherapeutic orientations. Finally, the response rate: at 49% this was a good return
rate for posted questionnaires but comparatively small in relation to the total population
of psychotherapists. It is probably the case that this study investigated 7.5% of the
approximately 2600 psychoanalytic psychotherapists and 10% of the 1097 BABCP
accredited/members who were clinical psychologists.
The epistemological measures had shortcomings that could affect the conclusions
drawn from this investigation. On the OMPI the Organicism dimension may be associated
256 Andrew R. Arthur
with more desirable, positive and humanitarian qualities than Mechanism, reported
Johnson and Miller (1990). They believe there are positive aspects to Mechanism that are
not being captured by the instrument. The OMPI is also vulnerable in another area. It
only has 6 strong items for measuring Mechanism vs. 16 for Organicism (Johnson &
Miller, 1990). It is possible, therefore, that the OMPI is vulnerable to a bias towards
Organicism. Finally, the OMPI may measure and reect some social-political ideologies.
The other epistemological measure, the PEP, was found by Johnson and Miller (1990)
to require radical revision for scoring. Their independent factor analysis raises some
questions concerning the viability of the original factors, therefore this study used both
original (Royce & Moss, 1980) and Johnson and Miller’s revised scoring (private
communication, Johnson, May 1997). A last point concerning the PEP was that in
this investigation a smaller subsample of participants took it, and this may affect
generalizability to the whole sample.
The quantitative methods and instruments used have assumptions concerning
personality structure and cognitive-epistemological traits. This study investigated
participants only within those parameters. Also the study was not able to clarify whether
those traits found to be affected by experience measured a tendency to change with age, or
whether they were measuring traits originally and currently associated with some senior
psychotherapists.
An area for further research is to investigate the predictive validity of these personality
and cognitive-epistemological patterns for orientation choice. Having established the
constellation of traits associated with qualiŽed cognitive-behavioural and psychoanalytic
psychotherapists, it should be possible to assess potential therapists and predict within
which orientation they might eventually practise. After training and a period of
experience the cohort could then be contacted to conŽrm their subsequent orientation.
It would be predicted from this study that those persons who showed evidence of an
objective-behavioural type proŽle would gravitate towards the cognitive-behavioural
therapies, and those with a subjective-analytic type proŽle would gravitate towards the
psychodynamic or psychoanalytical therapies.

A cknow ledgements
Professor Robert Bor, Department of Psychology, City University London—for supervising the research. Mr
Timothy Arthur and Dr Chris Barker, University College London, and Ms Doreen Carter—for assistance.
Alice Lovell, Lecturer, Southbank University and Dr Peter Scragg, Clinical Psychologist—for consultation.
Two anonymous reviewers for their helpful criticisms on an earlier version of this paper.

References
American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders (4th ed.).
Washington, DC: American Psychiatric Association.
Angelos, C. A. (1977). Relationships of psychotherapists’ personality and therapy methods (Doctoral
Dissertation, University of Michigan, 1977). Dissertation Abstracts International, 38, 1392B.
Arthur, A. R. (1998). An investigation into the personality traits and cognitive-epistemological styles of cognitive-
behavioural and psychoanalytic psychotherapists. Unpublished doctoral dissertation, City University, London.
Barron, J. (Ed.). (1978). The relationship between personality of the psychotherapist and his/her selection
of a theoretical orientation to psychotherapy [Special issue]. Psychotherapy: Theory, Research and Practice, 15
(4).
 Personality of psychoanalytic psychotherapists 257
Caine, T. M., & Smail, D. J. (1969). The effects of personality and training on attitudes to treatment:
Preliminary investigations. British Journal of Medical Psychology, 42, 277–282.
Costa, P. T., & McCrae, R. R. (1992). NEO P1-R, professional manual. Odessa, FL: Psychological Assessment
Resources Inc.
Cummings, N. A., & Lucchese, G. (1978). Adoption of a psychological orientation: The role of the
inadvertent. Psychotherapy: Theory, Research and Practice, 15, 323–328.
Hill, C. E., & O’Grady, K. E. (1985). List of therapist intentions illustrated in a case study and with
therapists of varying theoretical orientations. Journal of Counseling Psychology, 32, 3–22.
Johnson, J. A., Germer, C. K., Efran, J. S., & Overton, W. F. (1988). Personality as the basis for theoretical
predictions. Journal of Personality and Social Psychology, 55, 824–835.
Johnson, J. A., Howey, R. M., Reedy, Y. B., Gribble, H. A., & Ortiz, J. M. (1989, June). Extending the construct
validity of the Organicism-Mechanism Paradigm Inventory. Poster session presented at the Žrst annual
convention of the American Psychological Society, Virginia, USA.
Johnson, J. A., & Miller, M. L. (1990, November). Factor analysis of worldview inventories suggest two
fundamental ways of knowing. Paper presented at Fakultat fur Pedagogik der Universitat der Bundeswehr
Munchen, and Colloquium for Differential Psychology and Personality at Universitat Bielefeld.
Keinan, G., Almagor, M., & Ben Porath, Y. S. (1989). A reevaluation of the relationship between
psychotherapeutic orientation and perceived personality characteristics. Psychotherapy, 26, 218–226.
Kolevzon, M. S., Sowers-Hoag, K., & Hoffman, C. (1989). Selecting a family therapy model: The role of
personality attributes in eclectic practice. Journal of Marital and Family Therapy, 15, 249–257.
Kreitman, N. (1962). Psychiatric orientation: A study of attitudes among psychiatrists. Journal of Mental
Science, 108, 317–328.
Lazarus, A. A. (1978). Style not systems. Psychotherapy: Theory, Research and Practice, 15, 359–361.
Millon, T. (1994). Millon Index of Personality Styles-Manual. San Antonio: The Psychological Corporation
Harcourt, Brace and Company.
Royce, J. R. (1964). The encapsulated man. Princeton, NJ: Van Nostrand.
Royce, J. R., & Mos, L. P. (1980). Manual, Psycho-Epistemological ProŽle. Edmonton: Center for Advanced
Study in Theoretical Psychology, University of Alberta, Canada.
Scandell, J., Wlazelek, B. G., & Scandell, R. S. (1997). Personality of the therapist and theoretical
orientation. The Irish Journal of Psychology, 18, 413–418.
Schacht, T. E., & Black, D. A. (1985). Epistemological commitments of behavioral and psychoanalytic
therapists. Professional Psychology: Research and Practice, 16, 316–323.
Schwartz, B. D. (1978). The initial versus subsequent theoretical positions: does the psychotherapist’s
personality make a difference? Psychotherapy: Research and Practice, 15, 344–349.
Tremblay, J. M., Herron, W. G., & Schultz, C. L. (1986). Relation between therapeutic orientation and
personality in psychotherapists. Professional Psychology: Research and Practice, 17, 106–110.
Vasco, A. B., Garcia-Marques,L., & Dryden, W. (1993). ‘Psychotherapistknow thyself!’: Dissonance between
metatheoretical and personal values in psychotherapists of different theoretical orientations. Psychology
Research, 3(3), 181–196.
Walton, D. E. (1978). An exploratory study: Personality factors and theoretical orientations of therapists.
Psychotherapy: Theory, Research and Practice, 15, 390–395.
Walton, H. J. (1966). Difference between physically-minded and psychologically-minded medical practi-
tioners. British Journal of Psychiatry, 112, 1097–1102.
Weiss, L. G. (1997). The MIPS: gauging the dimensions of normality. In T. Millon (Ed.), The Millon
Inventories: Clinical and personality assessment (pp. 489–522). New York: Guilford.

Received 6 November 1998; revised version received 13 May 1999