Scandinavian Journal of Psychology, 2005, 46, 285–296
An empirical look at the Defense Mechanism Test (DMT): Reliability and
construct validity
Blackwell Publishing, Ltd.
BO EKEHAMMAR,1 IRENA ZUBER2 and MARJA-LIISA KONSTENIUS3
1
Department of Psychology, Uppsala University, and Stockholm School of Economics, Sweden
The Office of Cognitive Psychotherapy, Stockholm, Sweden
3
Department of Psychology, Stockholm University, Sweden
2
Ekehammar, B., Zuber, I. & Konstenius, M.-L. (2005). An empirical look at the Defense Mechanism Test (DMT): Reliability and construct
validity. Scandinavian Journal of Psychology, 46, 285–296.
Although the Defense Mechanism Test (DMT) has been in use for almost half a century, there are still quite contradictory views about whether
it is a reliable instrument, and if so, what it really measures. Thus, based on data from 39 female students, we first examined DMT inter-coder
reliability by analyzing the agreement among trained judges in their coding of the same DMT protocols. Second, we constructed a “parallel”
photographic picture that retained all structural characteristic of the original and analyzed DMT parallel-test reliability. Third, we examined
the construct validity of the DMT by (a) employing three self-report defense-mechanism inventories and analyzing the intercorrelations between
DMT defense scores and corresponding defenses in these instruments, (b) studying the relationships between DMT responses and scores on
trait and state anxiety, and (c) relating DMT-defense scores to measures of self-esteem. The main results showed that the DMT can be coded
with high reliability by trained coders, that the parallel-test reliability is unsatisfactory compared to traditional psychometric standards, that
there is a certain generalizability in the number of perceptual distortions that people display from one picture to another, and that the construct
validations provided meager empirical evidence for the conclusion that the DMT measures what it purports to measure, that is, psychological
defense mechanisms.
Key words: DMT, projective techniques, defense mechanisms, construct validity.
Bo Ekehammar, Department of Psychology, Uppsala University, Box 1225, SE-75142 Uppsala, Sweden . E-mail: bo.ekehammar@psyk.uu.se
INTRODUCTION
The Defense Mechanism Test (DMT) is a percept-genetic
(cf. Kragh & Smith, 1970) and projective test based on
assumedly anxiety-provoking TAT-like pictures, which are
exposed through a tachistoscope using gradually increasing
exposure duration ranging from 5 ms (subliminal) to 2,000
ms (supraliminal). The probably most often employed DMT
picture (cf. Olff, 1991) depicts in the centre a teenage person
(the Hero, in DMT terminology) of the same sex as the person
being tested, and in the periphery an older and threatening
person (the Peripheral person, in DMT terminology), of the
same sex as the Hero and the person being tested. This situation is meant to depict an Oedipal threat (an angry mother
or father) which is supposed to induce different types of
“precognitive defensive organization” (Kragh, 1962a) and
activate defense mechanisms characteristic of the perceiver.
The DMT was originally developed by Kragh (1955) and
has since then been employed for selection, especially in
military settings (e.g., Kragh, 1960, 1962b; Harsveld, 1991;
Neuman, 1978; Stoker, 1982; Stoll & Meier-Civelli, 1991;
Torjussen & Værnes, 1991; Værnes, 1982), and for diagnosis
in clinical contexts as well (e.g., Armelius, Sundbom,
Fransson & Kullgren, 1990; Bogren & Bogren, 1999; Fransson
& Sundbom, 1997; Jönsson, 1998; Rubino, Pezzarossa &
Grasso, 1991; Sundbom & Armelius, 1992; Sundbom, Binzer
& Kullgren, 1999). The rationale of the DMT as a selection
instrument for stressful occupations is that the psychological
defenses bind psychic energy necessary for coping with
stressful situations, and further, the defenses give rise to
perceptual distortions that make adequate perception in
dangerous situations difficult (cf. Kragh, 1985).
It should be emphasized that the DMT has been employed
in different ways as regards apparatus, stimulus pictures,
number of exposures and exposure times, instructions, followup questions, and evaluation of the DMT protocols (see
Olff, 1991, for a review). Further, whereas the original DMT
(cf., Kragh, 1985) is based on the classification of people’s
responses into various psychoanalytical defenses, some
recent attempts have been made to analyze DMT data without making such a classification but to base the analysis on
the empirical structure of naturally occurring response categories. Thus, Cooper and Kline (1989; see also Cooper, 1991)
have developed an “objectively scored” variant of the DMT
that rests on G-analysis to assess the similarity between
people’s response profiles as well as a Q-factor analysis of
these person similarity measures to obtain person factors
rather than factors based on correlations among variables.
In a similar vein, Armelius and Sundbom (1991, see also
Henningsson, Sundbom, Armelius & Erdberg, 2001) have
employed partial least squares principal component and discriminant analysis to map people into a multidimensional
space on the basis of their original DMT responses. In the
present study, however, we have used procedures in line with
the original methodology as expressed in Kragh’s (1985)
official DMT manual.
© 2005 The Scandinavian Psychological Associations/Blackwell Publishing Ltd. Published by Blackwell Publishing Ltd., 9600 Garsington
Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA. ISSN 0036-5564.
286
B. Ekehammar et al.
Although the DMT has been in use for almost half a
century, there are still conflicting opinions about its reliability and validity. For example, Harsveld (1991, p. 51) concluded that “the DMT is an extremely unreliable measuring
instrument” and Stoll and Meier-Civelli (1991) stated that
“the DMT as a whole is not convincing . . . there remain too
many open questions and uncertainties” (p. 94), and further,
“there are also problems with reliability of the test” (p. 95).
On the contrary, Cooper (1988, p. 381) stated that it is “the
most elegant and best-validated technique that exists for the
study of defense mechanisms in normal and neurotic
persons”, and Kline (1987) concluded that “the DMT is a
reliable test in a wide variety of samples. It is clear that
the DMT is a useful test both in clinical and occupational
practice, and as a purported measure of defenses it should
be valuable in theoretical research” (pp. 54–55).
However, Kline (1987, 1993) also points out that the
good predictive validity figures associated with the DMT,
obtained in Scandinavian military selection settings, do not
give any basis for the conclusion that DMT measures what
it is supposed to measure (“it works but for wrong reasons”). Further, some non-Scandinavian military selection
studies provide no empirical evidence at all for the DMT’s
predictive validity (e.g., Harsveld, 1991; Stoker, 1982; Stoll
& Meier-Civelli, 1991). Finally, Olff, Godaert, Brosschot,
Weiss and Ursin (1990, 1991) failed to obtain significant
correlations between DMT defenses and corresponding
defenses measured by some other methods.
In a previous study (Zuber & Ekehammar, 1997) based on
a non-clinical sample, we examined the effects of various
stimulus factors of the DMT by analyzing where the DMTdefense signs occur (distribution in exposure duration),
which part of the picture that is involved (distribution in
location), and which signs that co-occur (using correlation
and factor analyses). The results disclosed, among other
things, that the location of misperceptions to the central or
the peripheral person of the picture was the major explanatory principle for the distribution of “signs of defense” on
factors, rather than similarity according to some known
criteria of psychodynamic theory. Instead of capturing psychodynamic defense mechanisms, which is the theoretical
basis of the test, these results imply that DMT primarily
tends to measure misperceptions linked to the stimulus
picture and to general perceptual or information-processing
characteristics of the perceiver. Some critical remarks of this
study and its conclusions made by Kragh (1998, 2001) have
been responded to by Ekehammar and Zuber (1999).
In a more recent DMT study (Ekehammar, Zuber &
Simonsson Sarnecki, 2002), we tested the basic psychodynamic proposition of the DMT by having the original DMT
pictures redrawn without changing their structural properties, and in this way a friendly and a neutral DMT variant
were shaped. In contrast to predictions made from psychodynamic/DMT theory, that the threatful variant would
activate more “signs of defense” than the others, our results
Scand J Psychol 46 (2005)
disclosed that the three pictures activated the same amounts
of “signs of defense” and the same levels of various perceptual DMT thresholds. In line with the conclusions from our
previous study, these results suggest that rather than capturing psychodynamic defense mechanisms the DMT seems to
tap perceptual or information-processing difficulties in correct identification of brief stimulus exposures regardless of
their emotional contents. Thus, irrespective of the content of
a visual stimulus, people seem to differ in how quick or slow
they are in their visual information processing of briefly presented stimuli. This view has received further support by
some neuropsychological EEG studies by Eriksen, Nordby,
Olff and Ursin (2000) who showed that when exposed to
neutral sine-wave tones those classified as having high or low
DMT defense “differ in basic neurophysiological mechanisms connected with the way these subjects orient to new
stimuli” (p. 267). In conclusion, our interpretation is radically different from the psychoanalytical view on which the
DMT is based, which presupposes that the stimulus must be
anxiety-evoking in order to elicit psychological defenses
which in turn may cause perceptual distortions and misperceptions of the stimulus.
In the present study, our aim was to take a closer look at
the DMT’s reliability and still another look at its construct
validity. It is clear from the background and the citations
given there that there are quite contradictory views about
whether the DMT is a reliable measuring instrument, and if
so, what it really measures. Further, as expressed by Stoll
and Meier-Civelli (1991), among others, there is some doubt
about the positive DMT validity figures because most of
them have been obtained in studies carried out by persons
involved in the DMT; “The most serious charge that can be
raised against the DMT is the fact that good prognostic
validity has only been found by few scientists, such as Kragh
and Neuman as test authors and by Torjussen, Ursin and
Vaernes. In view of the other inconsistencies this does not
inspire confidence” (p. 95). Against the background of this
claim, it seems worthwhile to conduct an independent examination of DMT reliability and construct validity, and at the
same time provide confirmation or disconfirmation of the
findings and conclusions of our previous studies (Zuber &
Ekehammar, 1997; Ekehammar et al., 2002).
Thus, in the present study, we first analyzed the reliability
of the DMT by examining the agreement among three
DMT-trained psychologists in their coding of the same
DMT protocols (inter-coder reliability). Second, because a
test-retest approach to reliability estimation does not seem
feasible for an instrument like the DMT (if the picture has
been consciously perceived once, it is not regarded useful for
a re-test), we constructed a “parallel” photographic picture
(employing amateur actors) that retained all structural characteristics of the original (parallel-test reliability or generalizability). Third, we examined the construct validity of the
DMT by (1) employing three defense-mechanism instruments that, in contrast to the DMT, are known and used in
© 2005 The Scandinavian Psychological Associations/Blackwell Publishing Ltd.
DMT reliability and validity 287
Scand J Psychol 46 (2005)
an international context, and analyzing the intercorrelations
between DMT defense scores and corresponding defenses in
these instruments, (2) studying the relationships between
DMT defense scores and scores on trait and state anxiety,
and (3) relating DMT defense scores to measures of basic
and instrumental self-esteem.
The construct validations were based on fairly strict predictions that were deduced and formulated a priori for the
relationships between the DMT defense scores and the
scores on the other instruments employed. Thus, under (1)
above, we conducted a classical construct validation in line
with that of Olff et al. (1990, 1991) and predicted from our
information processing view that there would be no relationships between DMT defense scores (= perceptual distortions, according to our view) and scores on corresponding
defenses, or defenses of similar degree of maturity, obtained
through the three other instruments. If the DMT measures
defense mechanisms in a psychoanalytical sense (Kragh,
1985), the prediction would be that DMT defense scores
correlate with corresponding defense scores obtained from
the other instruments. Under (2) above, we predicted from
our viewpoint no systematic relations of DMT defenses (i.e.,
perceptual distortions) with state and trait anxiety whereas
the prediction from the DMT point of departure would be
that there are either positive or negative relationships
between DMT defenses and state and trait anxiety either
because people with high anxiety levels need psychological
defenses to reduce their anxiety (positive relation) or
because people with effective defenses can keep their anxiety
levels low (negative relation). Under (3) above, the information processing model predicts no systematic relations of
DMT total scores with basic and instrumental self-esteem or
their combination whereas the prediction from the defense
mechanism perspective would be that (a) the total DMT
defense score is negatively related to basic and positively
related to instrumental self-esteem (see Ihilevich & Gleser,
1986; Turkat, 1978), and (b) the combination of high basic
and low instrumental self-esteem (high genuine self-esteem
according to Turkat, 1978) would relate to a low total DMT
score whereas the combination of low basic and low instrumental self-esteem (low genuine self-esteem according to
Turkat, 1978) would relate to a high total DMT score.
METHOD
Participants
The participants were 39 female undergraduate psychology students
in the age range 20–43 years (M = 24.5 years) who took part in the
study voluntarily and in return for course credits.
Original and photographic picture
The original female version of the probably most often used DMT
picture (cf. Olff, 1991) was employed, showing a girl (Hero) in her
early teens sitting at a table in the centre of the picture. The girl has
a neutral look but behind the girl to the right is a middle-aged
woman (Peripheral person) with a grotesque and threatful look.
This TAT-like picture is drawn. A “parallel” photographic version
of this picture was arranged by letting two amateur actors, one
young and one middle-aged woman, “play” the same scene as in the
original DMT picture, with the same composition and approximately the same distances between Hero, Peripheral person and
table, as in the original picture. Several photo shots were taken and
that photograph was finally chosen which was judged by a small
group of researchers and psychologists (n = 5) to be most similar to
the original. In this way, the two pictures could be regarded to be
parallel as to the psychological situation and could thus be assumed
to trigger the same psychological defenses. On the other hand, the
pictures were different as to form (drawing vs. photograph) and
time (old-fashioned vs. modern) in order to avoid possible test-retest
effects.
General procedure
The participants were greeted on arrival by a female psychologist
trained in the DMT. Before the pictures were exposed, the participants were instructed to fill out a background information and a
state-anxiety inventory (see further below). Then, the pictures were
exposed to each participant individually and in accord with the
DMT manual (Kragh, 1985). However, the number of exposures
was 12 instead of 22 as suggested in the manual (see further below).
The testing procedure took on the average around 30 min. to complete. After this first session, the participants filled out the stateanxiety inventory again as well as some other self-report inventories
measuring trait anxiety, defense mechanisms, and self-esteem (see
further below). This phase also took around 30 min. to complete.
Then, the participants had a 5 min. pause in the same room after
which the second picture was exposed as above. The presentation
order of the original and photographic picture was balanced so that
in the first test session half of the participants were exposed to the
original and half to the photographic picture. The second test session took also around 30 min. to complete.
The DMT
Procedure. The two pictures were each exposed 12 times to each
participant using a tachistoscope especially constructed for DMT
use (see Kragh, 1985). The exposure duration was 5, 10, 17, 30, 55,
95, 165, 290, 500, 870, 1,150, and 2,000 ms, respectively. Thus, the
number of exposures was the same as in the Kragh (1962a) and
Cooper and Kline (1986) validation studies, but in contrast to their
studies, the exposure duration was gradually increased in line with
the DMT manual (Kragh, 1985). Further, in non-clinical samples,
the number of exposures can preferably be reduced (cf. Olff, 1991)
without losing important information but saving some time.
The instruction told the participants to try to verbalize and make
a drawing of what they had perceived after each exposure of the
picture. Follow-up questions were used in case of unclear responses
on the sex, age and mood of the person(s) perceived in the picture
(see Kragh, 1985, p. II:8; Olff, 1991, p. 157).
Defense categories. Ten different defense mechanisms were coded
according to the DMT manual (Kragh, 1985, p. IV:3):
1 Repression (e.g., the stimulus persons have a quality of rigidity
and inanimateness).
2 Isolation (e.g., hero and peripheral person are separated from
each other in the field; parts of the configuration are excluded).
3 Denial (existence of threat is denied explicitly).
© 2005 The Scandinavian Psychological Associations/Blackwell Publishing Ltd.
288
B. Ekehammar et al.
4 Reaction formation (the threat is turned into its opposite, a
positive relation between hero and peripheral person).
5 Identification with the aggressor (hero is the threatening person).
6 Introaggression (turning against the self, e.g., hero is sad, hurt or
worthless).
7 Introjection of opposite sex (incorrect sex of hero/peripheral
person).
8 Introjection of another object (incorrect age of hero/peripheral
person).
9 Projection (successive changes of hero before peripheral person
has become threatening).
10 Regression (the structure breaks down to a structure of an
earlier phase level).
Coding. The DMT protocols were coded and scored blindly in
accord with the DMT manual (Kragh, 1985) by a second psychologist trained in DMT scoring. For each defense category, an
unweighted sum of the number of signs of defense was computed as
well as a weighted sum. The weights were assessed after having
divided the number of exposures into three phases of equal length,
starting with the exposure where the participant could report something meaningful for the first time (P1). The number of signs in each
phase was multiplied with 1, 2 or 3 depending on their occurrence
in the early, middle or late phase, respectively. In addition, a sum of
the unweighted and weighted scores were computed across all
defense categories to make up a total “defense” score. Further, three
detection thresholds (measured in ms) were coded: P1 for the first
meaningful percept, T1 for the first time the threat was reported,
and C1 for the first correct perception of the major contents of the
picture (cf. Kragh, 1985).
Additional coders. To make possible estimates of the magnitude of
inter-coder reliability, another two coders (one researcher and one
psychologist) trained in the DMT procedure made independent and
blind codings of the protocols taken up by the first psychologist.
For this part of the study, 15 protocols from the original DMT
picture were randomly chosen for reliability estimation.
Defense mechanism inventories
Three different self-report defense mechanism inventories were
employed after translation to Swedish.
Defense Mechanism(s) Inventory (DMI). The DMI is the most
widely used self-report defense mechanism instrument (Davidson &
McGregor, 1998) and was originally developed by Gleser and
Ihilevich (1969). Reliability and validity support for the method has
been provided (Cooper & Kline, 1982; Cramer, 1988, 1991; Gleser
& Ihilevich, 1969; Gleser & Sachs, 1973; Ihilevich & Gleser, 1971,
1986, 1995; Juni, 1982; Vickers & Hervig, 1981; but see Blacha &
Fancher, 1977; Davidson & McGregor, 1998). The revised DMI
version (Ihilevich & Gleser, 1986) employed here purports to measure the relative intensity of five groups of psychological defenses
assessed in five different conflict areas (authority, independence,
competition, masculinity/femininity, situational) through ten brief
stories, two for each conflict area. Each story is followed by four
questions about the participant’s actual behavior, fantasy, thoughts,
and feelings, respectively. A forced-choice response format is
provided where five choices are given for each question, each corresponding to one of the five defenses being measured. The participant
is instructed to choose one alternative that is most representative,
and one that is least representative, of how he or she would react in
the situation. The five defense categories coded are Turning Against
Object (TAO; which includes identification with the aggressor and
displacement), Projection (PRO), Principalization (PRN; which
Scand J Psychol 46 (2005)
includes intellectualization, rationalization, and isolation), Turning
Against Self (TAS), and Reversal (REV; which includes denial,
negation, reaction formation, and repression).
Defense Style Questionnaire (DSQ). This instrument is based on
Vaillant’s (e.g., 1971) model of the hierarchy of defense mechanisms
derived from psychodynamic theory. The DSQ was originally developed by Bond, Gardner, Christian, and Sigal (1983) and has
received empirical verification from some validation studies (e.g.,
Bond, 1995; Bond & Vaillant, 1986; Bond, Perry & Gautier, 1989;
Pollock & Andrews, 1989; Vaillant, Bond & Vaillant, 1986) but not
from others (Perry & Hoglend, 1998; Spinhoven, van Gaalen &
Abraham, 1995; see also Davidson & McGregor, 1998). The revised
DSQ (Bond, 1986, 1991) used in the present study contains 88 items
that are constructed so as to catch the conscious derivatives of
various defense categories. The item statements are responded to on
9-point scales anchored by 1 (strongly disagree) and 9 (strongly
agree). The DSQ is focused on general defense styles rather than
single defense mechanisms and has been found to be useful to differentiate between adaptive and maladaptive defense styles in the
first hand (Bond, 1991). In the current use of the DSQ (e.g., Bond,
Parris & Zweig-Frank, 1994), responses to the items are summed
into the following defense styles: Factor 1 – Maladaptive Action
Defenses (which include acting out, passive-aggression, regression,
withdrawal, inhibition, projection); Factor 2 – Image-distorting
Defenses (which include splitting, omnipotence, primitive idealization); Factor 3 – Self-sacrificing Defenses (which include pseudoaltruism, reaction formation); and Factor 4 – Adaptive Defenses
(which include humour, suppression, sublimation).
Life Style Index (LSI). This instrument was originally developed
by Plutchik, Kellerman, and Conte (1979), and it is based on
Plutchik’s (e.g., 1980) theory of emotion. The LSI is built on the
conception (Plutchik, 1995; Plutchik et al., 1979) that different
emotions are linked to specific defense mechanisms, for example,
displacement is related to anger, compensation to sorrow, etc.
According to this theory, defense mechanisms are unconscious per
se but will be reflected in people’s reports of their emotions. An
elaborated circumplex model has been presented by Conte and
Plutchik (1993), where the relation between defense mechanisms,
emotions, and personality disorders is specified. The revised version
of LSI (Plutchik & Conte, 1989) was employed here, consisting of
97 items to which the participant has to respond whether the item
statement applies to him or her ( yes) or not (no). The LSI covers
eight defense categories: Denial, Reaction formation, Projection,
Displacement, Repression, Regression, Compensation, and Intellectualization. The empirical support for the LSI has been somewhat
contradictory as regards reliability and validity (cf. Conte &
Plutchik, 1993; Davidson & McGregor, 1998; Endresen, 1991; Olff
& Endresen, 1991; Plutchik et al., 1979).
State and trait anxiety scales
To examine participants’ degree of state anxiety, a new scale was
constructed for the present study. It was based on two previous
scales: (1) a Swedish version of the State-Trait Anxiety Inventory
(STAI; Spielberger, Gorsuch & Lushene, 1970) presented by Rosén
(1988) and comprising 20 items, and (2) the trait-anxiety scale of the
Karolinska Scale of Personality (KSP; Schalling, 1986), from which
17 items were selected and modified so as to express state instead of
trait anxiety. The 37 items of the state-anxiety inventory were
answered to on a 4-point scale, anchored by 1 (do not agree at all)
and 4 (agree completely). Scores were computed as unweighted sums
across items for the three subscales Somatic, Muscular, and Cognitive state anxiety as well as Total state anxiety.
© 2005 The Scandinavian Psychological Associations/Blackwell Publishing Ltd.
DMT reliability and validity 289
Scand J Psychol 46 (2005)
The participants’ trait anxiety level was measured using the 30item trait scale of the KSP (Schalling, 1986) having the same
response format as the state-anxiety scale above. Scores on the subscales Somatic (10 items), Muscular (10 items), and Cognitive (10
items) trait anxiety as well as Total trait anxiety were computed as
described for the state scale.
Self-esteem scales
Self-esteem was measured by an instrument containing two scales
developed by Forsman and Johnson (1996) and validated by the
same authors (Forsman & Johnson, 1996; Johnson, 1998; Johnson
& Forsman, 1995). The present version of the instrument comprised
31 items (statements) responded to on 5-point scales anchored by 1
(strongly disagree) and 5 (strongly agree). The two scales measure
Basic self-esteem and Instrumental (Earning) self-esteem, respectively, where basic self-esteem refers to a basic sense of self-worth or
trust that is independent of others’ opinions and reactions whereas
instrumental (earning) self-esteem means that the person is dependent on the confirmation from others to maintain his or her selfesteem. This distinction parallels the distinction made by Turkat
(1978) between genuine and defensive self-esteem.
RESULTS
Inter-coder reliability
Fifteen randomly selected DMT protocols were coded independently by three raters (A, B and C) and the reliabilities
were estimated based on the number of signs of defense
given by the raters on each defense category. The reliabilities
were computed as pairwise inter-coder correlations (productmoment coefficients), mean inter-coder correlation, Cronbach’s
α (expressing the reliability of the three coders’ average
scores), and intraclass correlation (expressing the expected
reliability of one rater’s scores). For two DMT defense categories (Identification with aggressor and Regression) no reliability estimates could be computed because of non-variance
in some of the coders’ scores. The estimated reliability
coefficients are presented in Table 1 which shows that with
one exception only (Isolation) the inter-coder reliabilities are
quite satisfactory. Thus, with one DMT coder, one could
expect to obtain reliabilities between 0.76 and 0.95 for all
defense codings except for Isolation, where a level of 0.55
could be expected.
Generalizability of DMT scores
Comparing the original and photographic DMT picture. Table 2
displays descriptive statistics for the DMT-defense scores
and total scores of the original and photographic picture
and t-tests for the difference in defense scores between these
pictures. As shown in Table 2 for both pictures, only three
of the DMT-defense categories are coded rather frequently
(Reaction formation, Introjection of opposite sex, Introjection of other object) whereas two are never or seldom used
(Identification with aggressor, Regression), and the others
are used infrequently. Of the single defense categories, three
(Reaction formation, Introjection of opposite sex, Introjection of other object) of the four significant differences
between pictures reveal that the photographic picture has
triggered more signs of defense than the original DMT picture. Further, the total defense score, unweighted as well as
weighted, shows a significantly higher level for the photographic than the original DMT.
One explanation for these differences, apart from differences in anxiety-eliciting characteristics of the two pictures,
might be found in the perceptual thresholds for each picture.
Thus, the mean thresholds were computed and tested for
significance by t-tests. The results revealed that for P1 (first
meaningful percept, in ms) there was a significant (p <
0.001) difference between the original (M = 55.6, SD = 49.3)
and photographic (M = 7.23, SD = 2.9) picture. Thus, the
photographic picture allowed a faster perception of something meaningful which means that more signs of defense
were possible to code for this picture across the remaining
exposures as compared to the original picture. However, the
other two thresholds (T1 and C1) showed no significant
mean differences between pictures (t = −1.80 and −0.40,
respectively) which means that perception of threat, and
correct perception, were similar for the two DMT-picture
variants.
Table 1. Inter-coder reliabilities for the DMT defense categories based on data from three coders (A, B and C) expressed as pairwise (r) and
mean (Mr) inter-coder correlations, Cronbach’s α, and intraclass correlations (ric )
Defense category
rAB
rAC
rBC
Mr
α
ric
1.
2.
3.
4.
6.
7.
8.
9.
0.83
0.34
0.94
0.77
0.79
1.00
0.94
0.78
0.81
0.92
0.94
0.79
0.81
0.98
0.97
0.78
0.67
0.50
1.00
0.81
0.88
0.98
0.97
1.00
0.77
0.58
0.96
0.79
0.83
0.99
0.96
0.85
0.91
0.78
0.97
0.92
0.92
0.99
0.98
0.95
0.76
0.55
0.91
0.78
0.92
0.89
0.95
0.86
Repression
Isolation
Denial
Reaction formation
Introaggression
Introjection of opposite sex
Introjection of another object
Projection
Note: Reliabilities for Defense no. 5 – Identification with aggressor and 10 – Regression could not be computed because of non-variance for
some coder.
© 2005 The Scandinavian Psychological Associations/Blackwell Publishing Ltd.
290
B. Ekehammar et al.
Scand J Psychol 46 (2005)
Table 2. Descriptive statistics for the original and the photographic DMT picture on each defense category and for total scores, and t-values for
the difference between pictures
Original picture
Photographic picture
Defense category
M
SD
M
SD
t (42)
1. Repression
2. Isolation
3. Denial
4. Reaction formation
5. Identification with aggressor
6. Introaggression
7. Introjection of opposite sex
8. Introjection of another object
9. Projection
10. Regression
Total, unweighted
Total, weighted
0.13
0.33
0.18
1.26
0.03
0.59
1.00
2.23
0.15
0.00
5.90
11.46
0.41
1.18
0.56
1.37
0.16
0.85
1.64
3.00
0.37
0.00
4.38
9.40
0.10
0.44
0.28
2.92
0.00
0.10
4.95
6.08
0.05
0.00
15.03
24.21
0.38
0.72
0.76
2.17
0.00
0.31
3.20
4.57
0.22
0.00
5.74
10.95
0.27
−0.52
−0.89
−5.13**
–
3.56*
−6.72**
−5.41**
1.67
–
−9.38**
−7.37**
* p < 0.01, ** p < 0.001.
A final analysis of the characteristics of the two pictures
was carried out by comparing their effects on state anxiety.
Thus, the difference in state-anxiety level before and after
picture presentation was compared using dependent t-tests.
The results disclosed that state anxiety, both general and
specific, was higher for both pictures before than after presentation. For somatic anxiety, the before-after difference was
statistically significant for both the original (t = 5.13, p <
0.001) and photographic (t = 3.99, p < 0.001) picture, which
was also the case for the total state anxiety score (t = 3.13,
p < 0.01 and t = 2.17, p < 0.05, respectively).
Parallel-test reliability. By correlating the participants’ DMT
scores on the original and photographic picture, an estimate
of parallel-test reliability was obtained for each defense
category and the total defense score (unweighted and
weighted values) as well as for the perceptual thresholds
(ms). The results are displayed in Table 3 and show low to
moderate figures. On four (repression, introaggression,
introjection of opposite sex, projection) of the eight defense
categories, there were no significant relationships between the
scores (weighted or unweighted) on the original and photographic picture. On two of the categories (isolation, denial),
there was a positive and significant relation for either the
unweighted or weighted scores, and on the remaining two
defenses (reaction formation, introjection of another object),
there was a positive and significant correlation for both the
unweighted and weighted scores (with r varying between
0.37 and 0.44). The total defense scores (summed across all
defense categories) disclosed a significant and fairly high
(r = 0.51) correlation for the weighted but not for the
unweighted scores. Finally, the various perceptual thresholds
did not display any significant relationships between the two
pictures. Looked upon as parallel-test reliabilities, none of
the figures given in Table 3 can be regarded satisfactory.
Table 3. Parallel-test reliabilities of the DMT defense categories,
total scores, and perceptual thresholds (P1, T1, C1) obtained through
correlations between data from the original and the photographic
DMT picture
Defense category/Threshold
Unweighted
Weighted
1. Repression
2. Isolation
3. Denial
4. Reaction formation
6. Introaggression
7. Introjection of opposite sex
8. Introjection of another object
9. Projection
Total DMT score
P1
T1
C1
−0.09
0.23
0.44**
0.44**
0.17
0.05
0.37*
0.22
0.30
0.27
−0.19
0.00
−0.09
0.36*
0.28
0.41**
0.15
0.29
0.42**
0.20
0.51**
Note: Reliabilities for Defense no. 5 – Identification with aggressor
and 10 – Regression could not be computed because of nonvariance in score distribution. P1 = threshold for first meaningful
percept, T1 = threshold for perception of threat, C1 = threshold for
correct perception of picture.
* p < 0.05; ** p < 0.01.
However, the weighted total score showed a certain generalizability from one picture to the other.
Construct validity
Correlations with scores from other defense-mechanism
instruments. The various defense instruments analyzed here
differ from each other in the number and character of
defense mechanisms. Thus, not all DMT defense categories
have an exact correspondence in the self-report instruments.
Table 4 gives an overview of a reasonable classification of
© 2005 The Scandinavian Psychological Associations/Blackwell Publishing Ltd.
DMT reliability and validity 291
Scand J Psychol 46 (2005)
Table 4. Self-report defense categories in the Life Style Index (LSI),
the Defense Mechanism Inventory (DMI), and the Defense Style
Questionnaire (DSQ) corresponding to DMT defense categories
Table 6. Correlations between DMT and Defense Style Questionnaire (DSQ) defenses
DSQ Factor
DMT
LSI
DMI
DSQ
Repression
Isolation
Denial
Reaction formation
Introaggression
Projection
Regression
Total
Repression
Intellectualization
Denial
Reaction formation
–
Projection
–
Total
–
PRN
REV
REV
TAS
PRO
–
–
–
Factor
–
Factor
Factor
Factor
Factor
Total
2
3
1
1
1
Notes: PRN = Principalization, REV = Reversal, TAS = Turning
Aggression against Self, PRO = Projection, Factor 1 = Maladaptive
action defenses, Factor 2 = Image-distorting defenses, Factor 3 =
Self-sacrificing defenses.
Table 5. Correlations between DMT defense categories and
corresponding (see Table 4) categories in the Life Style Index (LSI)
and the Defense Mechanism Inventory (DMI)
DMT
LSI
DMI
Repression
Isolation
Denial
Reaction formation
Introaggression
Projection
Total
0.01
0.09
−0.08
0.01
–
0.13
−0.03
–
−0.06
−0.03
0.18
0.25
−0.07
–
Note: No correlations are significant at p < 0.05.
which DMT-defense categories that correspond to which
defense categories in the self-report instruments (DMI, LSI,
DSQ). This classification was employed here to make possible an analysis of the convergent validity of the DMT. Table
5 shows the correlations (product-moment coefficients) of
the DMT defenses with the corresponding LSI and DMI
defenses. As displayed in the table, there were no significant
(p < 0.05) correlations for the six DMT defenses that have a
correspondence in the LSI or DMI. The largest correlation
coefficient (r = .25, p < 0.10) was obtained for the relation
between DMT and DMI Introaggression. The other ten
relationships that were possible to compute were all close to
zero and non-significant.
In Table 6, the correlations (product-moment coefficients)
between DMT and corresponding DSQ defenses are presented. In this case, because of a small number of corresponding categories, the defense categories have also been
divided into mature (DMT defenses Repression, Isolation,
Reaction formation; DSQ factors 1 and 2) and immature
(the rest of the DMT defenses; DSQ factor 3) on the basis
of psychodynamic theory (e.g., Kernberg, 1975). Thus, comparisons can be made between strictly corresponding categories as well as categories that should correspond on the basis
DMT defense
1
2
3
4
Repression
Isolation
Denial
Reaction formation
Introaggression
Introjection of opposite sex
Introjection of other object
Projection
Total
0.13
−0.08
−0.21
0.07
0.20
0.31*
0.08
0.04
0.18
−0.01
0.02
0.01
−0.06
0.23
0.14
−0.01
0.28
0.14
0.02
0.34*
0.00
−0.10
0.05
0.17
0.06
0.20
0.26
0.11
−0.11
−0.04
0.07
0.10
−0.11
0.07
−0.10
0.04
Notes: Italicized figures denote coefficients for corresponding
defenses, bold-face figures denote expected relations based on the
degree of maturity of the defenses.
* p < 0.05 (one-tailed).
of degree of maturity. As shown in Table 6, none of the three
correlations between strictly corresponding categories was
statistically significant. Further, only two of the ten expected
(provided that DMT measures defense mechanisms) relationships on the basis of degree of maturity displayed
significant (p < 0.05) outcomes. Finally, the DMT total
defense score did not show any significant relationship with
any of the DSQ defense factors.
Correlations with trait and state anxiety. The correlations of
DMT scores with trait and state anxiety are presented in
Table 7. As the subscores of the anxiety scales (somatic,
muscular, cognitive) displayed the same correlational picture
as the total anxiety scores, only the last-mentioned are given
in the table. The state-anxiety correlations are based on
those 20 participants who received the original DMT picture
during the first test session, before and after which stateanxiety data were collected.
As shown in Table 7, there was a positive but nonsignificant correlation of DMT total defense score with total
trait anxiety but a significant and fairly substantial relationship with total state anxiety, before (r = 0.52) as well as after
(r = 0.63) the DMT exposures. Further, of the eight correlations between the single DMT defenses and trait anxiety,
there were two significant coefficients, which is more than
expected by chance. Also, corresponding correlations for state
anxiety showed more significant coefficients than expected
by chance (three out of eight for the before and after measure, respectively). Thus, to sum up, one may conclude that
there were some significant correlations of fairly substantial
magnitude between DMT defense and state anxiety scores,
which lends some support to the view that the DMT measures defense mechanisms (but see Discussion section).
Correlations with self-esteem. The correlations (productmoment coefficients) of DMT defense scores with Basic and
© 2005 The Scandinavian Psychological Associations/Blackwell Publishing Ltd.
292
B. Ekehammar et al.
Scand J Psychol 46 (2005)
Table 7. Correlations of DMT defenses with total trait and state
anxiety (Before = before picture exposure, After = after picture
exposure)
State anxiety
Defense category
Trait
anxiety
Before
After
Repression
2. Isolation
3. Denial
4. Reaction formation
6. Introaggression
7. Introjection of opposite sex
8. Introjection of another object
Projection
Total score
0.29*
−0.09
−0.20
−0.02
0.11
0.31*
0.21
−0.05
0.25
0.37*
0.12
−0.43*
−0.17
0.16
0.28
0.53**
0.26
0.52**
0.33
0.05
−0.42*
−0.29
0.37
0.52**
0.54**
0.30
0.63**
Notes: Correlations for Defense no. 5 – Identification with
aggressor and 10 – Regression could not be computed because
of non-variance in score distribution.
* p < 0.05, ** p < 0.01 (one-tailed).
Table 8. Correlations of DMT defenses with basic and instrumental
self-esteem (SE)
Defense category
Basic SE
Instrumental SE
1. Repression
2. Isolation
3. Denial
4. Reaction formation
6. Introaggression
7. Introjection of opposite sex
8. Introjection of another object
Projection
Total score
−0.13
0.11
0.04
−0.11
0.19
−0.37*
−0.06
0.19
−0.18
0.34*
0.05
0.02
−0.03
−0.02
−0.06
−0.20
−0.11
−0.13
Notes: Correlations for Defense no. 5 – Identification with aggressor
and 10 – Regression could not be computed because of nonvariance in score distribution.
* p < 0.01 (one-tailed).
Instrumental (Earning) self-esteem are given in Table 8. Provided that DMT measures defense mechanisms, the expectation was that the total DMT defense score would be
negatively related to basic and positively related to instrumental self-esteem. However, the table shows that these two
correlations were negative and non-significant. Further, as
shown in Table 8, there were only two significant correlations between the single DMT defense categories and the
two types of self-esteem, which could hardly have been
predicted from the defense mechanism proposition. Rather,
they might have occurred by chance. Further, the combination of basic and instrumental self-esteem (after median
splits of the respective score distributions), which is the main
idea when using the present instrument (cf. Johnson &
Forsman, 1995), did not yield any significant outcome. Thus,
we must conclude that no meaningful relationships between
DMT and self-esteem scores were obtained.
DISCUSSION
The present study has examined various reliability and construct validity characteristics of the DMT. As to reliability,
one of our analyses focused on inter-coder reliability, that is,
the agreement among DMT-experienced coders’ ratings of
the same DMT protocols. The results disclosed that this
aspect of reliability was clearly satisfactory for all DMT
defenses with only one exception. Thus, with the exception
of the defense category Isolation, the reliability coefficients
fell within the interval reported by Kragh (1985) in the
DMT manual. However, this source does not give any clear
description of how the reliability indices have been computed (but see, e.g., Ozolins, 1989). Anyhow, it seems clear
that DMT-trained psychologists seem to achieve high agreement in their DMT ratings although it is quite obvious that
for some defense categories (Introjection of opposite sex =
incorrect sex of perceived person; Introjection of another
object = incorrect age of perceived person), the criteria for
coding are so objective that disagreement among judges is
almost impossible. The implication of the high inter-coder
reliability is that one basic aspect of DMT reliability is
satisfied which would make it possible to detect latent relationships between DMT and other constructs.
As test-retest reliability does not seem appropriate for a
projective test like the DMT (e.g., Kragh, 1985), we examined
parallel-test reliability instead. According to Kragh (1985,
p. III:1), parallel-test reliability estimation “presents a viable
alternative” to the test-retest method. “However, the pictures
of the first and of the second (third, etc.) perceptgenesis
should not be too similar, or we can expect the retest effect
to come into play” (Kragh, 1985, p. III:1). The reliability
coefficients for the eight single defenses that were possible
to compute in our study showed figures between −0.09 and
0.44 whereas the total defense scores showed a somewhat higher
agreement, 0.30 for the unweighted and 0.51 for the weighted
scores. Corresponding values for the perceptual thresholds
varied between −0.19 and 0.27. Looked upon as traditional
parallel-test reliability estimates, these figures are clearly unsatisfactory and below standard psychometric requirements. However, if we have succeeded in following Kragh’s (1985) advice,
that the versions “should not be too similar”, it can of course
be questioned if we have succeeded in creating a “parallel”
picture to the original in a strict psychometric sense. For
example, the number of perceptual distortions (“signs of
defense”) was substantially higher for the photographic
than the original picture, and the threshold for the first
meaningful percept was substantially lower. These findings
go together because an early report of a meaningful percept
in the series allows coding of more “signs of defense” (perceptual distortions) in the subsequent exposures. However,
differences between the pictures in these respects will not
necessarily affect the magnitude of the correlation coefficients which, for the reasons given, could perhaps be labeled
generalizability instead of reliability coefficients.
© 2005 The Scandinavian Psychological Associations/Blackwell Publishing Ltd.
DMT reliability and validity 293
Scand J Psychol 46 (2005)
Anyhow, there was a certain degree of agreement in the
participants’ total defense scores (number of misperceptions)
between pictures, especially when these were weighted on the
basis of where in the series (early or late) the misperceptions
occurred (late misperceptions are weighted more heavily than
early ones). Looking at previous DMT research, we could find
only one study (Harsveld, 1991) similar to ours on this point.
In a pilot-selection setting, Harsveld (1991) computed, among
other things, correlations between “signs of defense” for the
present (but male) original DMT picture and a second
(male) “parallel” picture included in the TAT package. The
correlations obtained between the two pictures for the various defense categories were very similar to our figures, varying between 0.06 and 0.31 (no coefficients for total defense
scores were computed), and Harsveld (1991) concluded that,
“all measures discussed thus far are characterized by poor
reliability. The mechanisms seem to be largely determined by
factors outside the personality of the testee . . .” (p. 60).
Our construct validation of the DMT defenses through
examining their relationships with corresponding defense
categories in three different defense-mechanism inventories
yielded rather discouraging outcomes for the defense mechanism hypothesis. First, there were no significant relations
between the total defense scores of the DMT and those of
the other instruments. Second, there were no significant relations between any of the single DMT defense scores and the
scores for corresponding defenses in the other three instruments. In fact, this outcome corresponds very well with the
results of previous DMT studies by Olff et al. (1990, 1991)
who examined the relationships between the DMT and two
of the self-report defense instruments (DMI and LSI) also
examined in our study: “The most striking finding of this
study however is that there are no substantial correlations
between the overall scores of the different defense questionnaires and the DMT. Also, there are no correlations between
the corresponding subscales of the defense measures in DMT
and those in the questionnaires” (Olff et al., 1991, p. 314).
Thus, our and Olff et al.’s (1990, 1991) examinations of the
construct validity of the DMT seem to converge on all
important points.
Although there appears to be no relations between people’s
responses to the DMT and their scores on various selfreport defense-mechanism questionnaires (which are correlated among themselves), it would be unfair to conclude that
this finding definitely invalidates the DMT. In accord with
classical psychometric thinking the conclusion would be that
the DMT lacks construct validity. But the situation is complicated by the fact that recent research in various psychological domains (e.g., memory, motives, attitudes, prejudice)
has made a distinction between explicit (slow, intentional,
conscious) and implicit (fast, automatic, unconscious)
aspects of the same phenomenon (see, e.g., Greenwald &
Banaji, 1995). Further, empirical studies on this issue show
that implicit and explicit measures of the same construct are
not necessarily associated. As just one example, and of relev-
ance in the present context, McClelland, Koestner, and
Weinberger (1989) used projective TAT pictures (supposed
to measure implicit processes) as well as self-report questionnaires (supposed to measure explicit processes) to study
need-for-achievement and power motives among people.
The results disclosed weak correlations between the implicit
and explicit measures of the same motive and showed that
the implicit and explicit motives were related to different
kinds of external variables as well. The study of psychological defense mechanisms might benefit from a similar
approach, that is, to examine the power of the implicit and
explicit measures, respectively, for predicting various types
of relevant external behaviors. In this way, a safer conclusion
could be drawn as to the relative utility of the supposedly
implicit (e.g., the DMT) and explicit (questionnaires)
approaches to the study of defense mechanisms.
It can be argued, however, that the processes involved in
DMT perception are not fully implicit as all the exposures
are not subliminal because of the increasing exposure duration up to 2,000 ms. Thus, quite a few DMT exposures are
definitely supraliminal for almost all people, which makes
possible cognitive elaboration of the stimulus and makes the
process explicit (slow, intentional, conscious) rather than
implicit. The conclusion is that the DMT is not easily
classified as measuring either explicit or implicit processes.
Then, one could argue that it is at least not possible to “fake
good” in the DMT because the test person cannot be
expected to know which answers are classified as defenses or
not. However, the situation is more or less the same as concerns some of the self-report defense mechanism instruments. In the DMI, for example, after having read a vignette,
the test person is instructed to answer questions about his or
her actual reaction, thought, fantasy, and feeling. Also in
this case, it seems improbable that the person could fake his
or her answers in some direction to get or avoid a score on
a specific defense mechanism. In conclusion, we mean that
the similarities between the DMT and the defense mechanism questionnaires, as discussed above, make it reasonable
to predict a certain correlation between the scores obtained
from them, provided that the DMT actually measures
defense mechanisms.
In the construct validation employing measures on state
and trait anxiety, we predicted no correlations whereas the
defense mechanism view would predict positive or negative
correlations of the participants’ total DMT defense scores
with their state and trait anxiety levels. The results disclosed
significant and fairly substantial relationships between DMT
and total state anxiety, before (r = 0.52) as well as after (r =
0.63) the DMT exposures but there were no significant relations with total trait anxiety. Thus, the results indicate that
people’s state-anxiety levels affect their DMT perceptions,
the higher their state anxiety the larger is the number of
perceptual distortions (“signs of defense”) in their DMT
responses. Further, this relation seems to hold for their state
anxiety observed before as well as after the DMT exposures.
© 2005 The Scandinavian Psychological Associations/Blackwell Publishing Ltd.
294
B. Ekehammar et al.
Thus, anticipatory anxiety or test anxiety seem to explain
this finding rather than the anxiety-evoking character of the
DMT picture.
The final construct validation examined the relationships
between DMT scores on the one hand and basic and instrumental self-esteem on the other. Our hypotheses were supported. Thus, people’s DMT defense scores do not seem to
be meaningfully related to their basic and instrumental selfesteem levels.
How do the present results fit into the picture obtained
from our previous studies (Ekehammar et al., 2002; Zuber
& Ekehammar, 1997) where we, in accord with some other
researchers (e.g., Cooper & Kline, 1989; Eriksen, Olff,
Mann, Sterman & Ursin, 1996; Eriksen et al., 2000; Stoll &
Meier-Civelli, 1991), questioned the proposition that perceptual distortions in the test are reflections of defense mechanisms in a psychoanalytical sense. Instead, we suggested that the
DMT seems to tap more general perceptual or informationprocessing difficulties in correct identification of brief
stimulus exposures regardless of their emotional contents.
First, the different construct validations sustained our previous view that the DMT does not catch defense mechanisms
in a psychoanalytical sense because the expected relationships with other defense mechanism instruments, with trait
anxiety, and with self-esteem did not show up. Second, there
was a certain generality in the participants’ total number of
perceptual distortions between the two pictures that we
employed, but a low generality for specific “defenses”, which
provides support to our contention that the DMT seems to
reflect more general information-processing difficulties when
trying to identify the content of brief stimulus exposures.
Thus, these results fit rather nicely into our previous picture.
In addition, the present results suggest that the perceiver’s
state anxiety level affects the DMT responses; the higher the
state anxiety the more perceptual distortions were observed.
As explained above, anticipatory or test anxiety is a more
plausible explanation to this finding rather than the anxietyevoking character of the DMT picture itself. That state
anxiety interferes with information-processing quality is a
well-known finding and might explain the relation between
state anxiety and perceptual distortions in the DMT.
As a final theoretical conclusion, we want to emphasize,
like we did in our previous DMT study (Ekehammar et al.,
2002), that the present findings do not contradict the view
that there exist unconscious psychological processes in general. However, in line with the conclusions of Kihlstrom
(1999) and others, our results suggest that the study of
unconscious processes using subliminal and other techniques do not support psychoanalytical theory. Thus, in a
review of experimental studies on the psychological unconscious, Kihlstrom (1999, p. 435) concluded that “(n)one
of the experiments reviewed involve sexual or aggressive contents, none of their results imply defensive acts of repression,
and none of their results support hermeneutic methods of
interpreting manifest contents in terms of latent contents.”
Scand J Psychol 46 (2005)
The study was supported by grants to Bo Ekehammar from the
Swedish Council for Research in the Humanities and Social Sciences (F617/94). The authors are obliged to Jeanette Wennergren
and Sophia Marongiu Ivarsson for administrating the DMT.
REFERENCES
Armelius, B.-Å. & Sundbom, E. (1991). Hard and soft models for
the assessment of personality organization by DMT. In M. Olff,
G. Godaert & H. Ursin (Eds.), Quantification of human defence
mechanisms (pp. 138–147). Berlin: Springer.
Armelius, B.-Å., Sundbom, E., Fransson, P. & Kullgren, G. (1990).
Personality organization defined by DMT and the Structural
Interview. Scandinavian Journal of Psychology, 33, 81–88.
Blacha, M. D. & Fancher, R. E. (1977). A content validity study of
the Defense Mechanism Inventory. Journal of Personality
Assessment, 41, 402– 404.
Bogren, L. & Bogren, I.-B. (1999). Early visual information processing and the Defence Mechanism Test in schizophrenia. Acta
Psychiatrica Scandinavica, 99, 460– 465.
Bond, M. (1986). Defense Style Questionnaire. In G. E. Vaillant
(Ed.), Empirical studies of ego mechanisms of defense (pp. 146–
152). Washington, DC: American Psychiatric Press.
Bond, M. (1991). Manual for the Defense Style Questionnaire.
Unpublished manuscript
Bond, M. (1995). Development and properties of the Defense Style
Questionnaire. In H. R. Conte & R. Plutchik (Eds.), Ego defense
theory and measurement (pp. 202–220). New York: Wiley.
Bond, M., Gardner, S. T., Christian, J. & Sigal, J. J. (1983). Empirical study of self-rated defense style. Archives of General Psychiatry, 40, 333 –338.
Bond, M., Paris, J. & Zweig-Frank, H. (1994). Defense styles and
borderline personality disorders. Journal of Personality Disorders, 8, 28–31.
Bond, M., Perry, J. C. & Gautier, M. (1989). Validating the
self-report of defense styles. Journal of Personality Disorders, 3,
101–112.
Bond, M. & Vaillant, J. S. (1989). An empirical study of the relationship between diagnosis and defense style. Archives of General Psychiatry, 43, 285–288.
Conte, H. R. & Plutchik, R. (1993). The measurement of ego
defenses in clinical research. In U. Hentschel, G. J. W. Smith,
W. Ehlers & J. G. Draguns (Eds.), The concept of defense mechanisms in contemporary psychology: Theoretical, research, and
clinical perspectives (pp. 275 –289). New York: Springer.
Cooper, C. (1988). The scientific status of the Defence Mechanism
Test. British Journal of Medical Psychology, 61, 381–384.
Cooper, C. (1991). G-analysis of the DMT. In M. Olff, G. Godaert
& H. Ursin (Eds.), Quantification of human defence mechanisms
(pp. 121–137). Berlin: Springer.
Cooper, C. & Kline, P. (1982). A validation of the Defense Mechanism
Inventory. British Journal of Medical Psychology, 55, 209–214.
Cooper, C. & Kline, P. (1986). An evaluation of the Defence Mechanism Test. British Journal of Psychology, 77, 19–31.
Cooper, C. & Kline, P. (1989). A new objectively scored version of
the Defence Mechanism Test. Scandinavian Journal of Psychology, 30, 228 –238.
Cramer, P. (1988). The Defense Mechanism Inventory: A review of
research and discussion of the scales. Journal of Personality
Assessment, 52, 142–164.
Cramer, P. (1991). The development of defense mechanisms: Theory,
research, and assessment. New York: Springer.
Davidson, K. & MacGregor, M. W. (1998). A critical appraisal of
self-report defense mechanism measures. Journal of Personality,
66, 965 –992.
© 2005 The Scandinavian Psychological Associations/Blackwell Publishing Ltd.
DMT reliability and validity 295
Scand J Psychol 46 (2005)
Ekehammar, B. & Zuber, I. (1999). In defence of the criticism of the
Defence Mechanism Test (DMT): A reply to Kragh (1998).
Scandinavian Journal of Psychology, 40, 85 – 87.
Ekehammar, B., Zuber, I. & Simonsson Sarnecki, M. (2002). The
Defence Mechanism Test (DMT) revisited: Experimental validation using threatening and non-threatening pictures. European
Journal of Personality, 16, 283 –294.
Endresen, I. M. (1991). A Norwegian translation of the Plutchik
questionnaire for psychological defense. Scandinavian Journal of
Psychology, 32, 105 –113.
Eriksen, H. R., Nordby, H., Olff, M. & Ursin, H. (2000). Effects of
psychological defense on processing of neutral stimuli indicated
by event-related potentials (ERPs). Scandinavian Journal of Psychology, 41, 263–267.
Eriksen, H. R., Olff, M., Mann, C., Sterman, M. B. & Ursin, H.
(1996). Psychological defense mechanisms and electroencephalographic arousal. Scandinavian Journal of Psychology, 37, 351–
361.
Forsman, L. & Johnson, M. (1996). Dimensionality and validity of
two scales measuring different aspects of self-esteem. Scandinavian Journal of Psychology, 37, 1–15.
Fransson, P. & Sundbom, E. (1997). Defense Mechanism Test
(DMT) and Kernberg’s theory of personality organization
related to adolescents in psychiatric care. Scandinavian Journal
of Psychology, 38, 95 –102.
Gleser, G. C. & Ihilevich, D. (1969). An objective instrument for
measuring defense mechanisms. Journal of Consulting and Clinical Psychology, 33, 51– 60.
Gleser, G. C. & Sacks, M. (1973). Ego defenses and reaction to
stress: A validation study of the Defense Mechanisms Inventory.
Journal of Consulting and Clinical Psychology, 40, 181–187.
Greenwald, A. G. & Banaji, M. R. (1995). Implicit social cognition:
Attitudes, self-esteem, and stereotypes. Psychological Review,
102, 4–27.
Harsveld, N. (1991). The Defence Mechanism Test and success in
flying training. In E. Farmer (Ed.), Stress and error in aviation
(Vol. 1, pp. 51–65). Aldershot, UK: Avebury Technical.
Henningsson, M., Sundbom, E., Armelius, B.-Å. & Erdberg, P.
(2001). PLS model building: A multivariate approach to
personality test data. Scandinavian Journal of Psychology, 42,
399–409.
Ihilievich, D. & Gleser, G. C. (1971). Relationship of defense mechanisms to field dependence-independence. Journal of Abnormal
Psychology, 77, 296 –302.
Ihilievich, D. & Gleser, G. C. (1986). Defense mechanisms: Their
classification, correlates, and measurement with the Defense Mechanisms Inventory. Owosso, MI: DMI Associates.
Ihilievich, D. & Gleser, G. C. (1995). The Defense Mechanisms
Inventory: Its development and clinical applications. In H. R.
Conte & R. Plutchik (Eds.), Ego defenses: Theory and development (pp. 221–246). New York: Wiley.
Johnson, M. (1998). Self-esteem stability: The importance of basic
self-esteem and competence strivings for the stability of global
self-esteem. European Journal of Personality, 12, 103–116.
Johnson, M. & Forsman, L. (1995). Competence strivings and selfesteem: An experimental study. Personality and Individual Differences, 19, 417– 430.
Jönsson, S. A. T. (1998). The Wechsler Adult Intelligence Scale
(WAIS), but not the Defence Mechanism Test (DMTm),
separates schizophrenics and normal controls in a factorial
cluster analysis. Scandinavian Journal of Psychology, 39, 109–
116.
Juni, S. (1982). The composite measure of the Defense Mechanism
Inventory. Journal of Research in Personality, 16, 193 –200.
Kernberg, O. (1975). Borderline conditions and pathological narcissism. New York: Aronson.
Kihlstrom, J. F. (1999). The psychological unconscious. In L. A.
Pervin & O. P. John (Eds.), Handbook of personality (2nd edn,
pp. 424 – 442). New York: Guilford Press.
Kline, P. (1987). The scientific status of the DMT. British Journal of
Medical Psychology, 60, 53 –59.
Kline, P. (1993). The Defence Mechanism Test in occupational psychology: A critical examination of its validity. European Review
of Applied Psychology, 43, 197–204.
Kragh, U. (1955). The actual-genetic model of perception-personality.
Lund, Sweden: Gleerup.
Kragh, U. (1960). The Defense Mechanism Test: A new method for
diagnosis and personnel selection. Journal of Applied Psychology,
44, 303 –309.
Kragh, U. (1962a). Precognitive defensive organisation with threatening and non-threatening peripheral stimuli. Scandinavian Journal
of Psychology, 3, 65 – 68.
Kragh, U. (1962b). Predictions of success of Danish attack divers
by the Defense Mechanism Test. Perceptual and Motor Skills,
15, 303 –309.
Kragh, U. (1985). The Defence Mechanism Test: Manual. Stockholm: Persona.
Kragh, U. (1998). In defence of the Defence Mechanism Test: A
reply. Scandinavian Journal of Psychology, 39, 123 –124.
Kragh, U. (2001). DMT interminable: A re-reply to I. Zuber and B.
Ekehammar. Scandinavian Journal of Psychology, 42, 135–136.
Kragh, U. & Smith, G. (1970). Percept-genetic analysis. Lund,
Sweden: Gleerup.
McClelland, D. C., Koestner, R. & Weinberger, J. (1989). How do
self-attributed and implicit motives differ? Psychological Review,
96, 690 –702.
Neuman, T. (1978). Dimensionering och validering av perceptgenesens försvarsmekanismer: En hierarkisk analys mot pilotens stressbeteende. [Dimension analysis and validation of percept-genetic
defence mechanisms: A hierarchical analysis of the pilot’s behaviour
under stress.] (Report no. C55020-H6). Stockholm: FOA [National
Defence Establishment, Division of Behavioural Sciences.]
Olff, M. (1991). The DMT method in Europe: State of the art. In
M. Olff, G. Godaert & H. Ursin (Eds.), Quantification of human
defence mechanisms (pp. 148 –171). Berlin: Springer.
Olff, M. & Endresen, I. M. (1991). The Dutch and the Norwegian
translations of the Plutchik questionnaire for psychological
defence. In M. Olff, G. Godaert & H. Ursin (Eds.), Quantification of human defence mechanisms (pp. 59 –71). Berlin: Springer.
Olff, M., Godaert, G., Brosschot, J. F., Weiss, K. E. & Ursin, H.
(1990). Tachistoscopic and questionnaire methods for the measurement of psychological defences. Scandinavian Journal of Psychology, 31, 89 –98.
Olff, M., Godaert, G., Brosschot, J. F., Weiss, K. E. & Ursin, H.
(1991). The Defence Mechanism Test related to questionnaire
methods for measurement of psychological defence. In M. Olff,
G. Godaert & H. Ursin (Eds.), Quantification of human defence
mechanisms (pp. 302–320). Berlin: Springer.
Ozolins, A. (1989). Defence patterns and non-communicative body
movements. Stockholm: Almquist & Wiksell International.
Perry, J. C. & Hoglend, P. (1998). Convergent and discriminant
validity of overall defensive functioning. Journal of Nervous and
Mental Disease, 186, 529 –535.
Plutchik, R. (1980). Emotion: A psychoevolutionary synthesis. New
York: Harper & Row.
Plutchik, R. (1995). A theory of ego defenses. In H. R. Conte & R.
Plutchik (Eds.), Ego defenses: Theory and measurement (pp. 13–
37). New York: Wiley.
Plutchik, R. & Conte, H. R. (1989). Measuring emotions and their
derivates: Personality traits, ego defenses, and coping styles. In
S. Wetzler & M. Katz (Eds.), Contemporary approaches to psychological assessment (pp. 241–269). New York: Brunner/Mazel.
© 2005 The Scandinavian Psychological Associations/Blackwell Publishing Ltd.
296
B. Ekehammar et al.
Plutchik, R., Kellerman, H. & Conte, H. R. (1979). A structural
theory of ego defenses and emotions. In C. E. Izard (Ed.),
Emotions and psychopathology (pp. 229 –257). New York: Plenum
Press.
Pollock, C. & Andrews, G. (1989). Defense styles associated with
specific anxiety disorders. American Journal of Psychiatry, 146,
1500–1502.
Rosén, A. (1988). State anxiety and abortion. Anxiety Research, 1,
115–125.
Rubino, A., Pezzarossa, B. & Grasso, S. (1991). DMT defences in
neurotic and somatically ill patients. In M. Olff, G. Godaert &
H. Ursin (Eds.), Quantification of human defence mechanisms
(pp. 207–221). Berlin: Springer.
Schalling, D. (1986). The development of the KSP inventory. In B.
af Klinteberg, D. Schalling & D. Magnusson, Individual development and adjustment (pp. 1– 8). Reports from the Department of
Psychology, Stockholm university, No. 64.
Spielberger, C., Gorsuch, R. & Lushene, R. (1970). The State-Trait
Anxiety Inventory (STAI): Test manual Form X. Palo Alto, CA:
Consulting Psychologists Press.
Spinhoven, P., van Gaalen, H. A. E. & Abraham, R. E. (1995). The
Defense Style Questionnaire: A psychometric examination.
Journal of Personality Disorders, 9, 124 –133.
Stoker, P. (1982). An empirical investigation of the predictive validity of the Defence Mechanism Test in the screening of fast-jet
pilots for the Royal Air Force. British Journal of Projective Psychology & Personality Study, 27, 7–12.
Stoll, F. & Meier-Civelli, U. (1991). An evaluative study of the
Defense Mechanism Test. In M. Olff, G. Godaert & H. Ursin
(Eds.), Quantification of human defence mechanisms (pp. 81–97).
Berlin: Springer.
Scand J Psychol 46 (2005)
Sundbom, E. & Armelius, B.-Å. (1992). Reactions to DMT as
related to psychotic and borderline personality organization.
Scandinavian Journal of Psychology, 33, 178 –188.
Sundbom, E., Binzer, M. & Kullgren, G. (1999). Psychological
defense strategies according to the Defense Mechanism Test
among patients with severe conversion disorder. Psychotherapy
Research, 9, 184 –198.
Torjussen, T. & Værnes, R. (1991). The use of the Defence Mechanism Test (DMT) in Norway for selection and stress research. In
M. Olff, G. Godaert & H. Ursin (Eds.), Quantification of human
defence mechanisms (pp. 172–206). Berlin: Springer.
Turkat, D. (1978). Self-esteem research: The role of defensiveness.
Psychological Record, 28, 129 –135.
Værnes, R. (1982). The DMT predicts inadequate performance
under stress. Scandinavian Journal of Psychology, 23, 37–43.
Vaillant, G. E. (1971). Theoretical hierarchy of adaptive ego mechanisms: A 30 year follow-up of 30 men selected for psychological
health. Archives of General Psychiatry, 24, 107–118.
Vaillant, G. E., Bond, M. & Vaillant, C. O. (1986). An empirically
validated hierarchy of defense mechanisms. Archives of General
Psychiatry, 43, 786 –794.
Vickers, R. R. & Hervig, L. K. (1981). Comparison of three psychological defense mechanism questionnaires. Journal of Personality
Assessment, 45, 630 – 680.
Zuber, I. & Ekehammar, B. (1997). An empirical look at the
Defence Mechanism Test (DMT): Stimulus effects. Scandinavian
Journal of Psychology, 38, 85 –94.
Received 18 June 2003, accepted 10 February 2004
© 2005 The Scandinavian Psychological Associations/Blackwell Publishing Ltd.