:Il... ;. .3!
Fit.

. _
inn». 5

. .
2125:: s A.

.:
,zmmmmﬁ
5: a",

,Luunv 5
i

. $31115}!
. 31:91:)
$3.5

3,331.31 .
I. {NSEQLXU
53::1:.£X<:.:;.l.( .
.y £35.94: at}?
. .. ‘
I. :liirui.zn.l .3...

t. ‘

{1235.3 :

7.. .5335.
. a .

..

9.5.3.24.

.5
r: P :5
if: s.

 

41.17... 5!...»- r1359
“Ail-.1. 1:1...

..§.z.a...
:siiﬁr‘t.
L£ug\h.i.l.z
: . .i
. .. In!

.5

1.15.5;33
.113: 2.5.

(«Eu

.. as if; .5
1329:. :vxbsls

\}.5)€.A.Il.zit it .51.

F. ilk—£5
S .

l .34.
I}

Li.}.4.n:ﬂ;§i.
i . 4 3:52..
auixwrﬁimlg

5.21..

$21.13;: 3.. :31:
tr..;..1..~.aa§’5
2.5«(115511

umittgiiiiirigiiiiiii

This is to certify that the

dissertation entitled

A VALIDITY STUDY OF THE MMPI-Z:
THE NON—GENERALIZABILITY OF THE ILLUSION OF
MENTAL HEALTH STUDY

presented by
Mark E. Heim

has been accepted towards fulﬁllment
ofthe requirements for

Ph.D. degree in Clinical Psychology

Mama Karon

Major professor

 

Date October 15, 1999

MSU i: an Afﬁrmuliw' Action/Equal Opportunity Insliluiion 0712771

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

DATE DUE DATE DUE DATE DUE
“02 to 300d;
0 7
1 3

 

6/01 c:/C|RC/DaieDue.965-p.15

A VALIDITY STUDY OF THE MMPI-Z:
THE NON-GENERALIZABILITY OF THE ILLUSION OF MENTAL HEALTH
STUDY
By
Mark E. Heim

Michigan State University

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements

for the degree of

DOCTOR OF PHILSOPHY
Department of Psychology

1999

ABSTRACT
A VALIDITY STUDY OF THE MMPI-2: THE NON-GENERALIZABILITY OF THE
ILLUSION OF MENTAL HEALTH STUDY
By
Mark E. Heim
The intent of this study was to extend the findings of the Illusion of Mental Health study
by Shedler, Mayman and Manis (1993), by examining the validity of various MMPI-2
scales regarding their ability to measure psychopathology. The Shedler et al. study cast
doubt on self-report measures to detect distress because they are susceptible to the
processes of denial. Thus, a distressed person using denial may appear healthy on self-
report measures. In that study they compared self-report assessments to both a clinical
impression of psychological health based on early memories, as well as physiological
indices. In the present study, to extend the results of Shedler et al.’s ﬁndings, 5 7
participants were administered the MMPI-2, the Defense Mechanism Inventory (DMI),
several cards from the Thematic Apperception Test (TAT) which were scored by
Westen’s (1990) Social Cognition and Object Relations Scale (SCORS), and heart rate
and blood pressure were recorded while undergoing various stressor tasks. There was
modest evidence that the MMPI-2 Welsh (1956) Anxiety scale can provide an accurate
assessment of psychological health when scores from the MMPI-2 Little and Fisher
(195 8) Denial scale are taken into account. This Denial scale, as well as the K scale were
correlated with an independent measure of denial on the DMI. Unexpectedly, the SCORS
data did not correlate with the physiological data of heart rate and blood pressure

reactivity to stress. The results indicated that the SCORS scales are susceptible to denial.

TABLE OF CONTENTS

List of Tables .................................................... v
List of Appendicies ............................................... vii
Introduction:

Introduction and Study Overview ..................................... 1
Hypotheses ..................................................... 6

Literature Review:

Defense Mechanisms and the Role of Denial ............................ 7
Problems in Deﬁning Denial ........................................ 10
The Measurement of Defense Mechanisms ............................. 17
Psychometric Properties of the DMI .................................. 20
An Overview of the MMPI and MMPI-2 ............................... 23
Development of the MMPI Validity Scales ............................. 26
An MMPI Derived Denial Scale ...................................... 37
MMPI-2 Measures of Global Psychological Health ....................... 43
Discussion of Marlowe Crowne Social Desirability Scale ................. 44
The Assessment of Psychological Health versus Distress .................. 44

The Use of Heart-Rate and Blood Pressure as Indices

of Defense and Distress ...................................... 49

 

 

Table of Contents (Con’t)

MM:

Participants ....................................................... 52
Instruments ....................................................... 52
Procedure ........................................................ 56
Scoring .......................................................... 58
knits:

Hypothesis 1 ....................................................... 60
Hypothesis H ...................................................... 61
Hypothesis III ..................................................... 64
Summary ......................................................... 64
Exploratory Analyses of the Study Measures of Psychological Health ......... 65
Unstructured Impressions of Global Psychological Health ................... 67

Summary of Exploratory Analyses between the SCORS and the
Psysiological Indices ......................................... 69

Exploratory Analyses Based on Group Classiﬁcations by MMPI-2 Scores

in Conjunction with Affect Tone Scores .......................... 69
Discussion:
Discussion ....................................................... 72
Conclusions ...................................................... 79
References ...................................................... 84

LIST OF TABLES

Table 1: Correlations between SCORS Assessment and Physiological Indices

of Psychological Health/Distress .......................... 92
Table 2: Correlations between MMPI-2 Subscales and Physiological Indices

of Psychological Health/Distress .......................... 93
Table 3: Correlations between MMPI-2 Scales of Psychological Health

and Physiological Indices of Psychological Health/Distress

While Controlling for Denial ............................. 94
Table 4: Correlations between MMPI-2 Subscales and SCORS Scales ......... 95
Table 5: Correlations between MMPI-2 Scales of Global Psychological Health

and SCORS Scales While Controlling for Denial ............. 96
Table 6: Mean Differences on Physiological Reactivity Scores in Healthy

versus Distressed Groups (based on a Mean Split of

Affect Tones scores) ..................................... 97
Table 7: Mean Differences on Physiological Reactivity Scores in Healthy,

Distressed, and Undetermined Groups

(based on Affect Tones scores) ............................ 98
Table 8: Mean Differences on Physiological Reactivity Scores in Healthy,

Distressed, and Illusory Groups (Based on Affect Tone

scores and Welsh Anxiety Scores) .......................... 99

List of Tables (Con’t)

Table 9: Mean Differences on Physiological Reactivity Scores in Healthy,
Distressed, and Illusory Groups (Based on Affect Tone scores
and Welsh Anxiety Scores) .............................. 100
Table 10: Mean Differences on Physiological Reactivity Scores in Healthy,
Distressed, and Illusory Groups (Based on the Combination of
Affect Tone and Emotional Intelligence scores, as well as
Welsh Anxiety Scores) .................................. 101
Table 11: Correlational Analyses of SCORS AT Scores

with Various Study Scales ............................... 102

Vi

Appendix A:
Appendix B:
Appendix C:
Appendix D:

Appendix E:

LIST OF APPENDICES

Defense Mechanism Inventory Example .................... 103
Little and Fisher Denial scale .............................. 104
Phrase Association Task .................................. 106
The Calculation of Socio-Economic Status ..................... 108

Global health Assessment Based on TAT Responses:

Explanatory Examples ............................. 109

Vii

A Validity Study of the MMPI-2:
The Non-Generalizability of the Illusion of Mental Health Study
Introduction and Study Overview

The validity of the MMPI-2, the most widely utilized personality test in the United
States (Lubin, Larsen, & Matarazzo, 1984), is in question as a result of a recent study by
Shedler, Mayman, and Manis (1993). Shedler et al. found that the assessment results of
self-report measures may be moderated by defensive denial. Denial, which is often a
signiﬁcant part of a person's psychological defense system, is deﬁned as the disavowal of
a thought, feeling, or situation (Levitt, 1980). Relevant to this study, denial is believed to
prevent a person from recognizing distress in their life, thus they are not likely to endorse
items on self-report measures that indicate distress.

Shedler et al.'s (1993) study found that a signiﬁcant portion of people "appeared"
psychologically healthy on self-report measures (i.e., Eysenck's Neuroticism scale, Beck's
Depression Inventory). However, when this group was assessed with physiological
criterion measures (heart rate and blood pressure reactivity) several people were
considered to be in psychological distress. They also evaluated these participants (i.e.,
those who looked healthy on self-report, and distressed on physiological indices) with a
sensitive clinical measure (the Mayman [1967] Early Memory Test, scored by Mayman)
and again found that several people were considered to be in psychological distress. The
early memory interpretations were considered mostly impervious to denial, and Shedler et
a1. argued that these measurements provided a more accurate or realistic assessment of
distress than self-report. When the self-report measure indicated psychological distress,

this was also indicated by the physiological measures and by the clinical measure. When

the self-report measure indicated low psychological distress and the early memory
interpretations indicated high psychological distress, then the physiological criterion
measures also indicated a high level of distress. Consequently, the self-report measure
was said to display an "illusion" of mental health in such cases. It should be noted that
Shedler et a1. do not state that self-report measures such as the MMPI-2 are necessarily
invalid, but instead suggest that such measures are probably valid if a person's score falls
in the distressed range, but the results are ambiguous if the score falls in the healthy
range-~where they are likely to be moderated by denial.

Shedler et a1. (1993) did not utilize the MMPI-2 in their study, however, they
believe that their results apply to it because many of the MMPI items are transparent, like
the self—report measures that they used in their study. Thus a person using denial is not
likely to endorse such items that may indicate psychological distress or pathology.
Shedler et al. also pointed out that the original MMPI healthy criterion group used to
establish norms was chosen without using adequate screening or assessment methods to
fully validate psychological health. As a result, many people in that criterion group likely
reported illusory health, as deﬁned above. However, when the MMPI was revised in
1989 a new criterion group was used to develop norms, and this group was subjected to
more rigorous screening. This included the collection of biographical information and
data from the Life Events Form and the Marital Adjustment Forms (Butcher, Graham,
Williams, & Ben-Porath 1990). Although the screening for the MMPI-2 was better than
the original MMPI screening, these data were nonetheless mostly self-report forms and

therefore can still be considered questionable.

These criticisms of the MMPI-2 and the results of Shedler et al.'s (1993) study
combine to cast serious doubt on the MMPI-2's ability to detect distress when the person
the person self-reports as healthy (i.e., the test may not reveal all of those people who are
truly distressed). However, Shedler et al. did not address the fact that the MMPI-2 has a
number of elements in it's inherent design that suggest that it should be able to detect
distress, even if a test-taker uses denial.

To begin with, although some MMPI-2 items are transparent in content, the
clinical scales on the MMPI-2 were derived from an empirical keying approach, in which
items were chosen based on how groups of people with known diagnoses responded to
them. As a result, the empirical keying approach reveals subtle items that discriminate
between different groups. As discussed by Meehl (1945), the signiﬁcance of such items
would not have been guessed by the test-maker and are subsequently equally mysterious
to the test-taker.

Also, Meehl and Hathaway (1946/1956) acknowledged that some people will
undoubtedly try to answer MMPI items to look socially desirable. However, by including
such items, you give people a good chance to distort their answers, which allows the
examiner to observe the extent to which this process occurs, which is the function of the
MMPI-2 validity scales. More speciﬁcally, the MMPI-2 contains 3 validity scales, one of
which--the K scale--is often considered a measure of defensiveness. The K scale was
speciﬁcally designed for the MMPI to detect when someone may be underreporting
distress, which would result in illusory health (Meehl & Hathaway, 1946/ 1956). The K
scale score is used to make corrections to a number of the clinical scales so that they will

more accurately reﬂect distress (Graham, 1993). Another MMPI-2 validity scale, the L

 

scale, was designed to pick up more blatant attempts to deny pathology. In addition to the
validity scales the MMPI-2 also contains a number of subscales that were developed to
assess denial (e.g., Haan 1977; Little & Fisher, 1958) and repression (e.g., Welsh, 1956),
one of which will be examined in this study.

Unfortunately, the MMPI-2 validity, clinical, and subscales all have equivocal
validity in the literature. This is partially due to the fact that many of the validity studies
are difﬁcult to interpret because the MMPI was compared against other self-report
measures. As established by Shedler et a1. (1993), if defensive denial moderates self-
report measure assessments, comparing two such measures with one another does not
establish their validity. Finding a good criterion measure has historically plagued validity
studies (e.g., Garb, 1998).

Despite the MMPI-2’s inconsistent validity ﬁndings and the criticisms levied
against it by Shedler et al. (1993), the number of studies that do support the MMPI-2's
validity, the built-in validity and clinical scales, as well as its extensive use, suggest that it
is a potentially sound instrument--though one that could beneﬁt from further validation
with suitable criterion instruments. To this end, the goal of this study is to assess the
MMPI-2's validity by determining its ability to detect psychological distress, even when
the test—taker may be relying on denial as a defensive style.

Speciﬁcally, MMPI-2 measures of global psychological health\distress and
defensiveness\denial, will be compared to both a clinical impression (based on Thematic
Apperception Test stories scored with Westen’s SCORS) and physiological indices (heart
rate and blood pressure) of psychological health\distress. This is similar to the

methodology that was used in the Shedler et al. (1993) study. Additionally, the

relationship between the MMPI-2 K scale and a denial scale will be compared with a
structured projective measure of defense mechanisms (the Defense Mechanisms
Inventory) to provide external support for their validity. These measures, and the rational
for their use, are described in greater detail below.

The Thematic Apperception Test will be reliably scored with a measure that
incorporates a clinical impression of health\distress. This assessment, as well as the
physiological assessments, will provide an index of psychological health\distress that is
mostly free of the processes of denial. These assessments will will be used in

correlational analyses with the MMPI-2 assessments of psychological health\distress.

 

Hypotheses

1.

It is hypothesized that the MMPI-2 K scale and an MMPI-2 derived measure of denial
(i.e., the Little & Fisher, 1958, Denial scale) will be meaningfully and signiﬁcantly
correlated (positively) to an independent measure of defensiveness/denial on the

Defense Mechanism Inventory.

. Based on the ﬁnding of the lack of speciﬁcity of self-report measures of health in the

Shedler et a1. (1993) study, it is predicted that the MMPI-2 will not provide an
accurate assessment of global psychological health\distress for those particpants that
appear healthy. This lack of speciﬁcity should be reﬂected in a lowered overall
correlation with two criterion indices of psychological health\distress (i.e., a clinical
impression and physiological measures of heart rate and blood pressure), unless
defensive denial is taken into account. This hypothesis will be conﬁrmed if the effect
of partialling out denial--as assessed by the MMPI-2 K scale or an MMPI-2 derived
scale of denial—signiﬁcantly increase the correlation between the MMPI-2
assessments of psychological health\distress and the criterion measures of
physiological reactivity as well as increase the correlation with scores on the Westen
SCORS scales.

It is hypothesized that as the measure of denial increases, the physiological reactivity

will increase, and the clinical impression of health based on the SCORS will decrease.

Literature Review

The following review will discuss defense mechanisms in general, and denial in
particular, followed by assessment techniques used to measure defense mechanisms. This
section is followed by a review of the MMPI, including the clinical, validity, and other
miscellaneous scales. Next, the review will focus on the assessment of psychological
health\distress using a TAT derived measure. The review will conclude with a discussion
on the use of physiological measures as indices of psychological health\distress.
Defense Mechanisms and the Role of Denial

In a review of psychoanalytic defense mechanisms, Sjoback (1973) reported that
Sigmund Freud began theorizing and writing about defensive processes in 1894 in an
essay titled "The Neuropsychoses of Defense". In these early writings, there were no
speciﬁc defense mechanisms and the concept of defense was primarily in relation to
drives. It was not until 1926, when S. Freud (1926/1959) published his monograph
Inhibitions, Symptoms and Anxiety, that his revised propositions about defense
mechanisms gained acceptance and recognition that have survived until today (e. g.,
Eagle, 1984; Sjoback, 1973, 1991).

According to Freud (1926\1959), the defensive process serves to protect the ego
from instinctual demands. In a chain-like sequence, a person experiences a forbidden
instinctual impulse which leads to a notion of danger, followed by painful affect and
subsequently defense. As a danger situation is experienced and anxiety is produced, the
objective of the defensive processes is to perform one or more of the following: block or
inhibit the mental content; distort the mental content; or screen and cover the mental

content by using opposite contents.

Thus, by the above actions a morally repugnant idea can be removed from
consciousness and the anxiety subsides. However, as discussed by Eagle (1984), most
defenses, except repression, allow a degree of discharge or gratiﬁcation of the instinctual
demand and thereby deﬂates part of the accumulated anxiety.

As discussed by Sjoback (1991), the concept of danger is central to S. F reud's
theory of defense, in that a forbidden or unruly impulse arouses a defense only if it
activates a notion of danger. Apart from an infantile experience of helplessness due to an
easily overwhelmed ego, S. Freud (1926/ 1 95 9) deﬁned four danger situations common in
early life: the loss of the object; the loss of the object's love; castration, and; punishment
from the superego.

Additionally, as the ego develops, the basic determinants of danger can undergo
modiﬁcations, thus infantile anxieties can retain an inﬂuence on adults (Sjoback, 1973).
Also noteworthy, the four basic dangers, and their modiﬁcations, have been found to be
frequent occurrences in any sample in any population (Sj oback,1993).

The concept of defense has been extensively expanded and written about since S.
Freud (e.g., Fairbairn, 1952/1994; Fenichel, 1945/1996; A. Freud, 1946; Kernberg, 1976),
and presently there is no agreed upon list of defense mechanisms (Plutchik, 1995). In
fact, Cramer (1991) reported more than 37 different defense mechanisms in the literature.

Of particular relevance to this study is the defense mechanism of denial. Denial is
generally agreed upon as being a central, common defense as evidenced by the number of
works that deal solely with this topic (e.g., Dorpat, 1985; Edelstein, Nathanson, & Stone,
1989). Additionally, Sjoback (1973) listed denial as one of the seven defense

mechanisms discussed by all early theorists, and it is relatively uncontested.

According to psychoanalytic thought, denial, often referred to and deﬁned as
disavowal (Levitt, 1980), is one of the earliest defense mechanisms available to a child.
As discussed by Cramer (1991), S. Freud's earliest theories on defense stated that, in
conjunction with repression, denial allows the warding off of upsetting perceptions of the
external world. Later theorists (e.g., Fenichel, 1945/ 1996) expanded the concept of denial
to include a warding off of internal stimuli as well. This process involved a covering
over, or a screen that substituted for painful thoughts.

Theoretically, denial is used to defend against unwanted and painful thoughts or
fantasies that are in the preconscious mind, whereas repression buries the thought or
feelings outside of awareness, in the unconscious (Cramer, 1991). Additionally, given
that denial involves a disavowal of external reality, in the extreme it has been considered
a forerunner, or possibly a manifestation of psychosis (Fine, 1975).

In a less extreme form, denial may be manifested by physically or psychologically
ignoring--or withdrawing attention from--a painful situation. In this circumstance, denial
is less absolute and it may be possible to lead a person in such a state to recognize what is
being ignored through sensitive questioning. In a less extreme form yet, denial may
operate through minimizing an event or ridiculing its importance (Cramer, 1991).

Eagle (1984) theorized about denial from a slightly different perspective. He
discussed the development of selﬂiood as an enabling to acknowledge thoughts,
memories, experiences, and wishes (i.e., one's engagement in the world) as one's own.
There is also the ability to fail to acknowledge, or disavowal, those engagements which
are inimical to the maintenance of selﬂiood and self-esteem (e.g., sexual and aggressive

thoughts and wishes). According to Eagle, such disavowal is considered the essence of

the psychoanalytic concept of defense, and he believes that this description is more
meaningful than describing the ensuing conﬂict that takes place when instinctual
impulses are pressing for discharge and ego structures are working to repress and control
them.
Problems in Deﬁning Denial

A common ﬁnding in much of the writing about defense mechanisms is that a
clear distinction is not made between denial and repression or suppression. Holmes
(1990) differentiated denial from repression as follows. Repression has 3 elements: (1)
selective forgetting of painful material; (2) it is not under voluntary control: and (3)
repressed material is not lost but stored in the unconscious and could be recalled if the
associated anxiety was resolved. The assertion that repressed material can not be
voluntary recalled is what differentiates repression from suppression and denial.

Singer and Sincoff (1990) discussed that deﬁning repression and defense is an
ongoing problem. Regarding repression, most writers use the term on one of 2 levels--as
a speciﬁc defense, or as a class of related defenses (which would include denial). Singer

and Sincoff, in their concluding chapter of Repression and Dissociation (edited by Singer)

 

have a subheading labeled "The Problem of Denial," which opens with the line "It is
intriguing that most of our contributors do not systematically address the distinction
between repression and denial." (p.483).

To further complicate matters, Dorpat (1985), in his book Denial and Defense,
uses the term denial as a general term for any reality-repudiating aspect of defensive
operations rather than as a discrete defense mechanism. Wegner and Pennebaker (1993)

uses the term "mental control" instead of denial, and their deﬁnition is as follows:

10

Psychological defenses are mechanisms of mental control that people use in
avoiding or manipulating mental states that they are strongly motivated to
inﬂuence. Mental control, in this view, is seen less as a willful choice and more
as an inevitable option toward which people are driven in the pursuit of relief

from anxiety, ego threat, or other psychic pain. (p. 4)

Vaillant (1990) provides an illustrative example which, from his perspective,
differentiates denial from suppression and repression, as well as a few other defense
mechanisms:

If a man were weeping in a cemetery but could not recall for whom he wept, this

would be repression. If he denied the existence of his tears, that would represent

psychotic denial. If he got the giggles or got drunk at the wake, that would be
dissociation (neurotic denial). If he said he wept from happiness, that would be
reaction formation. If he brushed aside his tears, said he would think about his
father's death tomorrow, and indeed remembered to grieve the next day, that

would be suppression. (pp. 262-263)

Fenichel (1945/ 1996) differentiated denial from repression using a developmental
perspective. He stated that repression becomes possible in a more developed ego than
was present when denial was used, however, repression is certainly a derivative of denial.
He further stated that the tendency to deny becomes weakened with the gradual
development of the ego and the strengthening of the reality principle with conjoint
memories and experiences.

According to Cramer (1991) many theorists agree that a major difference between

denial and repression is best understood from a developmental perspective, and there is a

11

body of research to support this perspective. For example, Cramer reviewed a numbered
of studies that found denial decreased with age, especially from about age 3 or 4 through
preadolescence. Thereafter, the use of denial, and other defense mechanisms, may wax
and wane throughout the life-span for a variety of reasons.

Shanan (1989) discussed that the use of denial, which may have tapered off over
the lifespan, may begin anew, or increase in older persons. He stated that older persons
are often required to disengage themselves from values and activities (e. g., driving,
increased social-isolation, impending death) due to physical or cognitive constraints. In a
similar manner akin to how regression may operate in the service of the ego, Shanan
stated that denial may enable a gradual disengagement from former commitments while
safeguarding the integrity of the ego. Shanan also discussed that as the use of denial
increases and more of reality is ignored, there is an increase in rigid and goal-directed
thinking, as well as an increase in fantasy activity--all of which have been empirically
shown to increase as a person ages.

In contrast to the developmental concept of various defense mechanisms, which is
often referred to as a horizontal approach, is the vertical or hierarchical approach to
defenses. From this second approach defenses are often classiﬁed according to some
dimension such as their degree of complexity or how much they distort reality (Cramer,
1991)

When conceptualizing defenses hierarchically, the issue is raised as to whether
denial, and defenses in general, are pathological or adaptive. As discussed in Cramer
(1991) and Sjoback (1973), S. Freud held to the view that all defense was pathological

and that neurosis and neurotic symptoms are related to the manifestations of different

12

defense mechanisms. (While S. Freud did not, in fact, think all uses of defense were
pathological, such a View is frequently attributed to him, since he mainly discussed
psychopathology rather than development). In contrast to this view, some contemporary
researchers believe that some defenses may lead to positive beneﬁts. For example, Taylor
& Brown (1988) suggest that certain "illusions" may be adaptive for mental health and
well-being. The middle ground in this debate is that defenses may serve a dual function.
For example, denial may hinder perception but it may also protect the ego from being
overwhelmed (Cramer, 1993)

Lazarus and Folkman (1991) discussed the issue of whether certain defensive
processes, which they refer to as "coping," are adaptive or destructive. They dismiss the
hierarchical View that is often discussed in the psychoanalytic literature, and instead
stated that:

Deﬁnitions of coping must include eﬂorts to manage stressful demands,

regardless of outcome. This means that no one strategy is considered inherently

better than any other. The goodness (efﬁcacy, appropriateness) of a strategy is
determined only by its effects in a given encounter and its effects in the long term.

. . . Predetermined ideas as to the inherent quality of ego processes prejudice us

against the possibilities of strategies ranked high in a hierarchy being maladaptive

and low-ranked strategies being adaptive. (p. 201)

Lazarus and Folkman (1991) used denial as an example to highlight how a low-
ranking defense may be viewed as adaptive. They acknowledge that the general
psychodynamic view of denial ranks it toward the bottom of a hierarchy, and it is

considered inherently maladaptive because of it's distortion of reality. Further, according

’13

to this view, although the use of denial in a stressful encounter may result in initial
emotional ease, over time there will be increased vulnerability to some deleterious
outcome.

For example, Schwartz (1990) found that denial in cardiac patients may enable
them to perceive or interpret their situation as safe, which has the effect of reducing
stress. However, if denial continues, the patient may ignore the seriousness of their
condition, which may impact their treatment or they may not engage in health promoting
behaviors. Thus, Schwartz stated that "Denial of illness becomes a gift in the short term
and a curse in the long term." (p.426).

However, a number of studies regarding the efﬁcacy of denial have found it to be
beneﬁcial. For example Shanan (1989) discussed the positive beneﬁts of denial in
situations of extreme, unalterable stress, such as a concentration camp or when facing a
life threatening illness.

Lazarus and F olkman (1991, p. 202) stated that " One can now speak of a box
score of studies with contradictory result, some showing that those who deny or avoid
threats are worse off than those who address them, and other studies in which denial is
associated with positive outcomes." (See Lazarus and Folkman, 1991, pp. 205-206, for a
listing of these studies; see also Edelstein, Nathanson, & Stone, 1989, for several more
examples of positive and negative aspects of denial).

As an aside, these contradictory ﬁndings discussed in Lazarus and F olkman
(1991), regarding whether denial is beneﬁcial or not, raises two important issues, one of
which they addressed. First, as mentioned above, there is a deﬁnitional problem in

studies that are purportedly assessing denial--usually deﬁned as the disavowal of reality.

’14

Lazarus and Folkman stated that several studies on denial may actually be better
conceptualized as avoidance, or I would add suppression. For example, an advanced
cancer patient may not talk about his condition and keep it out of mind as much as
possible, but he may readily acknowledge his reality and distress if confronted. Further,
other coping processes that have been classiﬁed as denial may be better conceptualized as
positive thinking, or minimization, which allows sustained motivation, morale, and
constructive efforts to cope--not necessarily a disavowal of reality (Lazarus & Folkman,
1991).

The second issue, not addressed by Lazarus and F olkman (1991), which may shed
light on some of the contradictory ﬁndings of studies on denial is a point made by Shedler
et al. (1993). That is, the results of many studies that relied solely on self-report
instruments to assess health\distress may be suspect because, as discussed at the outset of
this study, their validity has been called into question. A case in point was discussed by
Temoshok (1993), who discovered apparently contradictory results in two of her own
studies that looked at emotional expression and distress in patients when they were
informed they had malignant cancers, and how that affected their long-term outcome.
Temoshok was able to resolve the contradiction when she realized that one study used
clinical interviews to rate expression and distress, whereas the second study used self-
report measures, such as the Beck Depression Inventory. (It would be interesting to
review the contradictory studies listed in Lazarus and Folkman, 1991 , pp. 205-206 to see
what sort of outcome measures were used and how that may have inﬂuenced the results).

Returning now to Lazarus and Folkman's (1991) proposition that denial should

not be viewed as an inherently low level defense mechanism or coping strategy, they

15

offered several principles that they believe are a better method by which to evaluate
denial as being either adaptive or destructive. For example, denial may be adaptive to
alleviate stress when a person is faced with a threatening situation in which there is no
action they can take. Also, denial may be adaptive in the early stages of a crisis, when the
entirety of the situation would be too overwhelming to deal with all at once. Thus,
according to Lazarus and Folkman, a defense mechanism must be evaluated in context to
determine whether or not it is adaptive. Breznitz (1983) also discussed that denial is not
inherently maladaptive, and that to get an accurate assessment of it's efﬁcacy one must
look at what other alternatives are available to the individual (see also a relevant
discussion by Haan, 1977, pp. 175-179).

Regarding the deﬁnitional problem of denial as discussed by Lazarus and
Folkman (1991, see above), recognizing that denial may be either adaptive or maladaptive
depending on the situation or context in which it is used may help explain the controversy
between "positive illusions" described by Taylor (Taylor & Brown, 1988, 1994) and the
"illusion of mental health" (Shedler et al., 1993, 1994). In this controversy Taylor
maintains that a distortion of reality that results in overly positive self-evaluations and
unrealistic optimism promote mental health, whereas Shedler et a1. maintain that
underlying this self-deception, a careful clinical assessment will often ﬁnd distress.

In summary, denial is one of many defense mechanisms and is closely related to
suppression and repression. The main difference being whether or not the unwanted
thought, feeling, or fantasy can be easily acknowledged, as in suppression, or can not be
acknowledged or recalled, as in repression, with denial falling between the two. Thus,

the unwanted thought, feeling, or fantasy may be conscious, preconscious, or unconscious

16

and the corresponding defense would then be labeled suppression, denial, or repression,
respectively. However, given the current lack of speciﬁcity in the literature regarding
these defensive processes and their realm of operation, it seems necessary to view the
processes more on a continuum than as discrete ﬁinctions. This is consistent with a
current trend to also view the unconscious, preconscious, and conscious along a
continuum (Weinberger, 1990).

For the purposes of this study, denial shall be deﬁned from a general perspective,
not as a discreet, independent defense mechanism. That is, denial includes to some extent
what would be thought of as defensiveness at one end of the continuum, and reaches into
the area of repression on the other end. This is necessary because, as discussed below,
there is no valid assessment technique to solely measure denial. Rather, the best current
assessment techniques can only tap if a person is using a cluster of defenses which
resembles denial. Unfortunately, more ﬁne grained differentiations must remain in the
theoretical realm. Additionally, denial in this study is viewed as a low-level defense, with
deleterious physiological consequences. These consequences are discussed later in this
review.

The Measurement of Defense Mechanisms

After a review of the literature Cramer (1991) reported encountering 58 different
measures for defense mechanisms. The Rorschach has been one of the more common
tests used to assess defenses and is presently ﬂourishing, and to a much lesser extent the
Thematic Apperception Test (Sjoback, 1991). However, the majority of tests used to
assess defenses have been self-report measures--few of which survive for any length of

time (Sjoback, 1991). Apart from projective and self-report measures, two other common

17

modes of assessing defenses are (1) via vignettes that then require the test-taker to report
or chose how he would respond, and (2) via an interview method.

After reviewing a number of measures discussed in review articles and books
speciﬁc to defense mechanism assessment (Conte & Plutchik, 1995; Cramer 1988, 1991;
and Sj oback, 1991), as well as an independent review of the literature in the area of
defenses, the Defense Mechanisms Inventory (DMI; Gleser & Ihilevich, 1969; Ihilevich
& Gleser, 1986) appears to have the best reliability and validity and thus was chosen for
the this study. The DMI has been frequently used (Ihilevich & Gleser, 1986, contains a
bibliography of nearly 200 studies using the DMI), it assesses a cluster of defenses akin to
denial, it is easy to use and score, and it has acceptable psychometric properties. These
issues are further addressed below, after a description of the DMI.

The DMI consists of ten vignettes that assess 5 defense clusters. After each story
the test-taker is asked to chose from ﬁve responses which one most likely (scored 2), and
least likely scored 0), resembles how they "actually" would respond to the situation. The
remaining 3 choices are each scored 1. This format is repeated 3 more times for each
story except that the test-taker is subsequently asked how he would "impulsively"
respond, what "though " might occur to him, and how he "feels" about it. The rational for
beginning with overt behavior was that by allowing the test-taker to express actual
behavior, the effects of socially desirable responding might be reduced for the following 3
responses. An example of a DMI story and subsequent responses is contained in
Appendix A. There are male and female story versions for adults, children, and older
persons. The following information on the DMI is taken from Ihilevich and Gleser

(1986), unless otherwise noted.

’18

The 10 vignettes were conceptualized to represent 5 areas of conﬂict, with 2
vignettes for each of the 5 areas. The conﬂicts represented are: situational conﬂicts (e. g.,
physical safety needs pitted against unpredictable events), authority conﬂicts (e. g., self-
assertion may result in punishment), independence conﬂicts (e. g., satisfaction of needs is
pitted against the risk of deprivation of needs), competition conﬂicts (e. g., ambition and
status seeking are threatened by failure or humiliation), and either masculinity (for males)
or femininity (for females) (e. g., assertion of sexuality is threatened by sexual rejection).

To ascertain if the vignettes reﬂected the conﬂict situations, 5 judges with clinical
experience were asked to match the stories with the appropriate conﬂicts. Four of 5
judges successfully matched all the stories (the ﬁfth judge misread the instructions for
rating the stories which may explain his or her lack of agreement).

From the vignettes, 5 defense clusters are assessed: Turning Against the Object
(TAO); Projection (PRO); Principalization (PRN); Turning against the Self (TAS), and;
Reversal (REV). Theses are described below.

Turning Against the Object (TAO)

TAO involves the expression of direct or indirect aggression as a means to control

a perceived external threat. This cluster best represents the defenses of

identiﬁcation with the aggressor and displacement.
Projection (PRO)

PRO is the strategy of attributing negative intent or characteristics to others,

without evidence for doing so. Then the person who uses this strategy feels

justiﬁed in the expression of hostile thoughts, behavior, and feelings toward the

other person.

19

Principalization (PRN)
PRN entails a falsiﬁcation of reality by reinterpreting it through truisms,
platitudes, or sophistry. There is a sense that genuine understanding is replaced
by empty truth. This cluster best represents the defenses of intellectualization,
rationalization, and isolation.
Turning Against the Self (TAS)
TAS includes intrapunitive maneuvers that are deployed to falsify reality and
reduce perceived threats to self-esteem. The maneuvers include exaggerated self-
criticism, negative expectations, and depressed affect that work to preserve self-
esteem in a paradoxical fashion. No traditional defenses are mentioned under this
cluster.
Reversal (REV)
REV entails a minimization of the severity of perceived threats or conﬂicts and\or
a failure to acknowledge the existence of obvious dangers. This cluster includes
the defenses of denial, negation, repression, and reaction-formation.
Psychometric Properties of the DMI
Ihilevich and Gleser (1986, 1995) reported that 3 of the defense clusters showed
good content validity when each of the 240 DMI responses were examined by 10 clinical
judges unaware of the DMI clusters. The judges were asked to name a classical defense
mechanism represented by each response and their answers were compared to the DMI
defense clusters. For PRN, TAS, and REV clusters there was 60%, 80%, and 70%
matching agreement, respectively. PRO and TAO fared less well, showing only 45% and

43%, matching agreement, respectively, and several of PRO and TAO responses were not

20

considered defensive. Ihilevich and Gleser reported that subsequent studies replicated
these ﬁndings.

A weakness of the DMI reported in several studies is intercorrelations on some of
the defense clusters. Speciﬁcally, TAO and PRO reveal an average r = .45, and REV and
PRN reveal an average I: .48 (reported in Cramer, 1991). Cramer (1991) offered as an
explanation of these ﬁndings that either some of the DMI items have poor content
validity, or that defense clusters may overlap in reality. However, despite these
intercorrelations, both Ihilevich and Gleser (1995), and Cramer (1991), note that there is
considerable evidence that the scales are assessing independent dimensions. For
example, even though REV and PRN are intercorrelated, research has revealed that while
REV was related to denial, low dream recall, few postoperative complaints and successful
therapy, PRN was not related to these variables (Cramer, 1991). Other examples of
independence between REV and PRN can be found in the literature (e. g., Peglar &
Borgen, 1984).

Cramer (1991) reported that it is difﬁcult to establish concurrent validity because
there is a lack of other comparable defense measures to use as a criterion (this issue will
also be discussed below as it applies to certain MMPI subscales). However, when using
what measures were available, Cramer reported that REV consistently showed the
expected relationships with criterion measures of denial, primitive defense, repression,
and avoidance (no speciﬁc detail or analyses were reported). He further reported that
there was moderate criterion validity support for PRO, TAO and TAS, and poor support

for PRN.

21

Sjoback (1991) also reviewed the DMI, and his major contention was that the
vignettes are not adequate to evoke anxiety and thus elicit defenses as deﬁned in the
psychoanalytic tradition. However, at least one study (Gleser & Sacks, 1973) has shown
that a person's DMI proﬁle is predictive of their defenses used in an actual threatening
situation. Cooper and Kline (1982) also conducted a validation study of the DMI and
reported that it was consistent with proﬁles on the 16PF, which were generated based on
Freudian theory.

As discussed earlier, denial may have short term beneﬁts but be detrimental in the
long run. A study that lends validity to the REV scale that is consistent with this view
was conducted by Peglar and Borgen (1984). They followed 73 male patients that were
admitted to a coronary care unit and followed for ﬁve years, at which time 25 of the
subjects had died. A cluster analysis of distinctive DMI patterns revealed that those
subjects who scored high on PRN and low on REV were most likely to be alive at the end
of the ﬁve year period (high REV alone was associated with a positive perception of
health). Those subjects who scored high on both PRN and REV were least likely to be
alive at the end of the study. Other studies are consistent with the notion that denial, in
the long run, has detrimental results. For example, Minsky (cited in Ihilevich & Gleser,
1986) found that excessive reliance on denial (i.e., high REV score) was associated with
increased blood pressure.

Some studies are more difﬁcult to interpret because they relied on self-report
methods of assessment, which as discussed above, may be questionable (Shedler et al.,
1993). For example, Clum and Clum (1973) reported that subjects who used denial as a

defense mechanism self-reported less depression than those who used other defense

22

mechanisms. A study by Wilson (1982) reported that female postoperative patients who
had scored high on the REV scale fared better after surgery than those who scored low.
However, this study relied on a mixture of self-assessment and objective measures (e. g.,
amount of pain medicine requested) that make it more difﬁcult to interpret.

The DMI REV scale will be used in this study despite the shortcomings addressed
above, because it appears to be one of the best measures of denial available in an area
where valid measures are consistently difﬁcult to establish. This is important to note
because REV will be used as one of the criterion measures of denial against which certain
MMPI-2 scales will be compared. As a reminder, one purpose of this study is to ascertain
the ability of various scales on MMPI-2 to measure denial. As will be discussed in the
next section, the MMPI-2 K scale and the MMPI-2 derived Denial scale (Little & Fisher,
195 8) are two promising scales that are under investigation in this study. These scales are
mentioned now to note that REV is signiﬁcantly correlated with the K scale in a group of
college students for both males and females. In that same sample, REV is also
signiﬁcantly correlated to the MMPI-2 Hysteria scale for males (but not for females),
which is the scale from which Little and Fisher derived their MMPI-2 Denial scale
(Gordon & Brackney, 1979; see also Ihilevich and Gleser, 1986 for similar results).

An Overview of the MMPI and MMPI-2

In the late 1930's Starke Hathaway and John McKinley, while working at the
University of Minnesota Hospitals, were aware that 30 to 70 percent of patients seeing a
doctor for medical problems had what turned out to be psychoneurotic problems.
Subsequently, they set out to devise a test that could be used for routine diagnostic

assessments, and their efforts resulted in the MMPI, ﬁrst published in 1943. McKinley

23

and Hathaway hoped that their paper-and-pencil test would be a more efﬁcient and
reliable way of diagnosis--eliminating the need for individual interviews, mental status
exams, and individual psychological testing (Colligan, 1985; Graham, 1993).

The MMPI and its various scales were developed with an empirical keying
approach. Speciﬁcally, items that distinguished different groups of subjects with known
diagnoses became keyed as representative of each respective group. Most tests used at
that time were based on logical keying and questions that seemed to have face validity
based on the developer's subjective judgment of items reﬂecting whatever attributes were
being measured. The logical keying approach became increasingly unsatisfactory because
it was apparent that subjects could falsify or distort their responses to present themselves
however they desired. Further, research began accumulating on the logically keyed tests
which revealed inconsistencies between the self-report data and what was actually
observed between different groups of subjects. It was because of this research, and the
observations mentioned above, that the MMPI empirical keying approach was deveIOped
(Graham, 1993).

To develop the MMPI items, Hathaway and McKinley collected approximately
1,000 personality-type statements from existing psychological tests, textbooks, research
publications, and clinical case history experience. After deleting repeated items and
subsequent review and modiﬁcations, the MMPI consisted of 550 items that covered 26
different symptom categories. Originally, the items were printed on cards that the test-
taker would place into piles of true, false, or cannot say, as they applied to his or her self-

assessment (Colligan, 1985).

24

After item selection, criterion groups were chosen so that individual scores could
be meaningful compared and interpreted. The normal group consisted of relatives and
visitors of patients in the Minnesota University hospitals, as well as other groups that the
university had access to (e. g., high school graduates attending precollege conferences,
Graham, 1993). People in these groups were assumed normal and were not screened with
any other psychological instruments or interviews.

The second criterion group was made up of patients at the university hospital
diagnosed with a broad range of psychiatric problems. Discrete groups were formed
based on diagnostic labels used at that time, and if there was doubt at to the correct
diagnosis or if multiple diagnosis were present, the patient was not included. Eight
criterion groups were formed representing: hypochondriasis, depression, hysteria,
psychopathic deviation, paranoia, psychasthenia, schizophrenia, and hypomania. These
groups represented 8 clinical scales used on the MMPI (two categories--Masculinity-
Femininity, Social Introversion--were added later) (Graham, 1993).

To construct the scales, each criterion group's responses to the MMPI were
analyzed to identify the items that signiﬁcantly differentiated the speciﬁc criterion group,
other clinical groups, and the group of normals. Those items that resulted in
discrimination were then included in the resulting MMPI scale for that group.
Additionally, other steps were taken to cross-validate the clinical scales, and an attempt
was made to differentially weight some the items which did not pan out. (see Colligan,
1985; Graham, 1993).

The major criticisms levied against the MMPI included high intercorrelations of

many of the clinical scales, serious concerns about the original standardization sample,

25

offensive and\or archaic content in some items, and questionable diagnostic accuracy and
breadth of psychological problems assessed in the criterion groups (e.g., Anastasi, 1988;
Graham, 1993). These concerns, with the exception of interscale correlations, have been
mostly resolved when the MMPI was revised in 1989 to the MMPI-2 (Graham, 1993;
Hathaway and McKinley, 1989). Fortunately, by design, there are major similarities
between the two versions, thus it has been suggested that most of the research carried out
on the MMPI is applicable to the MMPI-2 (e. g., Graham, 1993; Hathaway & McKinley,
1989; Levitt, 1989).

Development of the MMPI Validity Scales

 

Because self-report measures at the time of the MMPI's introduction were
developed by intuitive methods and were shown to be susceptible to faking, Hathaway
and McKinley included four validity scales to detect deviant test-taking attitudes. First,
the Cannot Say scale provides the number of items not completed by a test-taker. If too
many items are omitted, the clinical scale scores will be lower, and the overall
interpretability of the test is questionable (Colligan, 1985).

Second, the L scale, also referred to as the Lie scale, was designed to detect
unsophisticated or naive attempts of test-takers to present themselves in a favorable light.

This is achieved by including items that require an admission to minor weaknesses in
character or personality common to most people (Colligan, 1983; Graham, 1993).

The next validity scale designed was the F scale to detect random endorsement of
items. A group of items was selected that was always answered in the same direction by
90 percent of the normal reference group. Thus if a person endorses a high number of

deviant responses on the items that make up this scale, the test proﬁle is considered

26

suspect (Colligan, 1983; Graham, 1993).

Finally, the K-scale, constructed by Meehl and Hathaway (1 946/ 1956) was
considered the most complex validity scale, and was intended as a measure of
defensiveness. In their classic paper, Meehl and Hathaway began their discussion by
stating:

One of the most important failings of almost all structured personality tests is their

susceptibility to "faking" or "lying" in one way or another, as well as their even

greater susceptibility to unconscious self-deception and role-playing on the part of
individuals who may be consciously quite honest and sincere in their responses.

(p. l 2)

The K scale was developed to address this issue, which echoes the concerns
discussed by Shedler et al. (1993) regarding self-report instruments, including the MMPI.

Meehl and Hathaway (1946/1956) discussed different attempts used on
personality measures to reckon with the problem of faking. One idea was to use "subtle"
items whose signiﬁcance are not apparent to the test-maker or the test-taker, yet the items
are for some reason able to discriminate criterion groups. Another way of dealing with
the problem is to acknowledge faking as a source of error, then attempt to correct for it.
More speciﬁc to the development of the K scale, test proﬁles that revealed no marked
pathology from individuals with known psychiatric problems (referred to as test "misses"
were compared to test proﬁles from psychiatric patients that did reveal pathology.

Meehl and Hathaway (1946/1956) reported undergoing several attempts
comparing test proﬁles as just discussed. They summarized that "the resultant scales

were about equally effective and about equally unsatisfactory regardless of the approach

27

and of the particular item content" (p.23). They worked an additional two years testing
various scales on different criterion groups which resulted in a scale for tentative adoption
originally referred to as L6.

According to Meehl and Hathaway (1946/1956), the content of the items in this
scale reﬂect "an attitude of denying worries, inferiority feelings, and psychiatrically
unhealthy symptoms, together with a disposition to see only good in others as well as
oneself. The over-all impression is one of 'impunitiveness'" (p.24). Additionally,
whereas a high score suggests defensiveness, a low score suggests plus-getting or
malingering (i.e., false positives).

Additional testing with the L6 scale on depressed and schizophrenic patients
resulted in the inclusion of 8 additional items, for a total of 30, and subsequently the new
scale was named K. Although Meehl and Hathaway (1 946/ 195 6) hinted that the K scale
score might be interpreted itself as a clinical scale, its real function was to provide
correction of the other scores. Eventually, the K scale was used to differentially correct
the scores of some scales (i.e., Hypochondriasis, Psychopathic Deviate, Psychasthenia,
Schizophrenia, Hypomania), while leaving the others uncorrected, which seemed to
provided the most accurate prediction of a person's clinical status (McKinley, Hathaway,
& Meehl, 1956).

Meehl and Hathaway (1946/1956) provided validity support for the new K scale
when they used a simple cutting score on the K scale to arbitrarily predict 44 randomly
selected MMPI proﬁles as either coming from normal subjects (if K 5 49) or psychiatric
subjects (if K _>_ 50). The 44 proﬁles came from 22 subjects previously classiﬁed as

psychologically healthy, and 22 psychiatric hospitalized patients. Based on the K score

28

alone, 37 of the MMPI proﬁles were correctly classiﬁed as coming from either the
healthy or abnormal group--an 85% hit rate. Six of the 7 errors resulted in classifying the
truly healthy as abnormal (i.e., false positives). This last ﬁnding is inconsistent to what
would be expected from Shedler et al.'s (1993) work, discussed in the introduction.

A shortcoming of the K scale is uncertainty regarding if and when correction
scores should be applied to a test proﬁle. That is, it is unknown when correcting for K
enhances, or is detrimental to, clinical accuracy or prediction. This was true on the
original MMPI (Dahlstrom, Welsh, and Dahlstrom, 1972; Heilbrun, 1961), and continues
to be an issue on the MMPI-2 (Graham, 1993).

Graham (1993) recommended that K-corrected scores be routinely used, given
that interpretations on the clinical scales are based on corrected scores. However, he
acknowledged that there were no empirical studies on the MMPI-2 to support the efﬁcacy
of using k-correction. As alluded to by Meehl and Hathaway (1956), beyond being used
solely for correction, the K scale has routinely become clinically interpreted as a measure
of defensiveness (e.g., Graham, 1993; Heilbrun, 1961; Smith, 1959). However, Graham
(1993, p. 30) pointed out that although high K scores typically indicate defensiveness,
moderate elevations sometimes reﬂect ego strength and psychological resources. He
continued to state that there is no deﬁnitive way to differentiate the two, and suggested
that if the test-taker appears psychologically healthy and functions reasonably well, then
the K score is probably reﬂecting more positive characteristics than defensiveness.

At least two studies (Heilbrun, 1961; Smith, 1959) that used the K scale as a
measure of defensiveness concluded that the K scale does an adequate job of measuring

defensiveness in abnormal populations, but both studies suggested that high K scores in

29

 

normal populations reﬂect psychological health (Heilbrun’s ﬁnding was speciﬁc to
females). However, the methodology and determination of healthy versus not healthy is
highly suspect. For example, in the Heilbrun (1961) study, subjects considered not
healthy were students that requested vocational or educational counseling. It is unclear
how health was rated in the Smith (1959) study, but it appears to be based on how the
subjects fared on a task that was supposed to reﬂect insight.

More recently, a meta-analysis was conducted by Baer, Wetter, and Berry (1992)
to assess the effectiveness of the MMPI K scale (as well as other scales) in detecting
underreporting of psychopathology. Their results revealed that, for underreporters, the K
scale was elevated by an effect size of nearly 1 standard deviation. However, their
deﬁnition of "underreporting" included defensiveness, as well as faking-good, and
socially desirable responding. In fact, of 25 studies included in the analysis, the
methodology of 17 studies simply gave normal subjects the instruction to try to appear as
well-adjusted as possible and these proﬁles were compared to normal subjects instructed
to respond honestly.

The major difﬁculty in properly examining the K scale for defensiveness is the
lack of an external criterion standard (recall that this issue was also raised in regard to
measuring defense mechanisms). Nearly every study of defensiveness, as assessed by the
K scale, involved subjects given instructions to fake good, or, subjects were used that
were assumed to have real-life reasons to fake-good (e. g., child custody hearings), but no
independent criterion conﬁrmation of defensiveness was carried out (Dahlstrom, Welsh,

& Dahlstrom, 1975).

30

To address this speciﬁc problem Lanyon and Lutz (1984) designed a study that
they believed included subjects that were likely to be in denial. In their study they
investigated 90 males arrested for felonious sexual offenses (mostly child molestation).
Based on police investigations which almost always resulted in conviction, the men
arrested were assumed guilty and the researchers classiﬁed the men as: no denial if the
subject account is the same as police account (n = 48), part denial if there was some
admission to sexually deviant behavior (n = 24) and, full denial if there was no admission
to sexually deviant behavior (11 = 18). Each subject completed the MMPI as part of an
assessment to determine rehabilitation potential or insanity/competency. The study
revealed few differences between the full denial group and the part denial group, thus
they were combined and compared to the no denial group. This analysis revealed
signiﬁcant differences beyond the .01 level on the L, F, and K scales as well as 3 derived
validity indexes being tested: L + K, L + K - F, and F - K. A subsequent multiple
regression revealed L + K - F to be the best predictor variable for denial versus no denial
group status.

Another avenue to be explored to help determine if the MMPI contains adequate
measures to assess denial or defensiveness is research which used physiological measures
or other health variables. Dahlstrom, Welsh, and Dahlstrom (1975, pp. 91-93)
summarized several studies of MMPI investigations using autonomic measures. Early
investigations unsuccessfully tried to determine if responses to certain test items tapped
test distortion (i.e., nonverdical answers), by monitoring subjects on a lie detector (using
GSR measures) while they took the MMPI. The results tended to show large variations

on different items.

31

Another unsuccessful method to determine if certain items might be indicative of
distortion is reported in Dahlstrom et a1. (1975) that involved monitoring the test-takers
response time (RT) to MMPI items, hoping to ﬁnd items that evoked long RT's which
would be inferred as defensiveness. Again, there was considerable intersubj ect variation.

Blumberg, West, and Ellis (1956) studied personality characteristics of
individuals with established malignant diseases hoping to ﬁnd individual differences
when following disease activity and survival. Several instruments were administered to
these individuals including the MMPI, the Rorschach, the TAT, and an intelligence
measure. They found the MMPI showed the most striking differences between the slow
versus rapid disease progressing groups. Speciﬁcally, those with rapid progression of
disease tended to have highly negative F-K values, which was considered indicative of
high defensiveness.

In contrast to the above study, Persky, Kempthome-Rawson, and Shekelle (1987)
reported that the only MMPI scale associated with cancer incidence and mortality in 2018
men was the depression scale. Their data did not support their hypothesis that repression,
as measured by the Welsh R scale, was associated with cancer risk.

There are a number of studies in the literature that attempted to ﬁnd MMPI
differences in survival rates for heart patients after surgery. Lair and King (1976)
reported conﬂicting ﬁndings in these studies (e. g., one study they cited found survivors
had signiﬁcantly higher denial based on L and K scores, while another study found
female nonsurvivors to have increased denial). In their own study Lair and King found

no signiﬁcant differences on the MMPI between male survivors and nonsurvivors, but in

32

the female nonsurvival group they reported higher MMPI scores on a scales of
Hypochondriasis and Hysteria.

In a separate study, Pancheri et a1. (1978) reported MMPI data on 58 male
intensive care heart patients. Their ﬁndings revealed that the group judged least
improved (based on medical parameters such as the occurrence of serious cardiac
arrhythmia) 7-10 days after admission had higher scores on almost all the MMPI scales.
However, the L and K scores were lower in the less improved group (L signiﬁcantly
lower, K not signiﬁcant). This ﬁnding may be interpreted such that increased
defensiveness and/or denial (based on higher L and K scores) leads to salutary effects, at
least initially.

However, as discussed earlier, Schwartz (1990), while reviewing data on cardiac
patient survival factors, pointed out that denial may be beneﬁcial in the short term but
deleterious in the long run. The studies reviewed so far focused on cardiac patient
survival or status shortly after hospitalization.

Tappan and Weybrew (1982) conducted a study that compared MMPI proﬁles of
1015 male submariners with their calculated coronary heart disease (CHD) risk (based on
7 parameters such as blood pressure, cholesterol level, electrocardiographic
abnormalities). The CHD risk data were compared with the 13 MMPI scales as well as
all possible combinations of subtests in pairs, triads, or tetrads. The results revealed 6
MMPI combinations (1 dyad, 3 triads, and 2 tetrads) of MMPI proﬁles that were
correlated with increased risk for CHD, all of which contained the K scale. The Hy and
Hs scale were also included in several of the MMPI combinations. Thirty-eight MMPI

combinations were associated with the low CHD risk group--none of which contained the

33

K, Hysteria, or Hypochondriasis scales. Unfortunately, this study did not include any
follow up data.

Another study was conducted by Gillum, Leon, Kamp, and Becerra-Aldama
(1980) to determine if personality factors as measured from the MMPI would be
associated with CHD. This study followed 281 men over a 30 year period and tested
them at 4 points. At the end of the 30 years, 62 men died from some type of cardiac
disease. Several different analysis were performed on the MMPI data which did not
reveal any predictive power for the age of onset of CHD or life expectancy.

Another issue that has surrounded the K scale, as well as the MMPI in general, is
that some researchers (e.g., Edwards, 1957; Fordyce, 1956; Fricke, 1956, Hanley, 1957)
claimed that people responded to personality tests with certain "response styles" or "sets."

One of these sets especially relevant to the present investigation was a social desirability
response set (SDRS). The researchers that adhered to the SDRS hypothesis interpreted
test results in a manner that assumed that a large proportion of people evaluated and
responded to questions on the MMPI to appear as socially desirable as possible.

To support this hypothesis, Edwards (1957) revealed that measures of social
desirability (based on his MMPI derived scale) correlated extremely high with many
scales on the MMPI (Approximately .70 for the K scale). Fordyce (1956) reported that
Edwards' social desirability scale correlated highly with the MMPI F and K scales, and
higher with each of them than they do each other. He interpreted this ﬁnding as an

indication that the common factor in the validity scales is best conceptualized as social

desirability.

34

 

To refute the SDRS hypothesis, Block (1965) discussed that Edwards’ (1957)
social desirability scale contained items that, if agreed with, would reﬂect a susceptibility
to anxiety (e.g., bodily tension, personal vulnerability). In fact, Edwards' 39 item scale
contains 22 items from the Taylor Manifest Anxiety (T MA) scale (Taylor, cited in Block,
1965), though keyed in the opposite direction. Block pointed out that, given this item
overlap, the two scales have to be correlated at about -.7, and the TMA is known to
correlate with several MMPI scales.

Block (1965) also discussed that it is nearly impossible to conceive of any
manifestations of neuroticism that would not be evaluated as socially undesirable.
Therefore there has to be an intrinsic relationship between expressions of personality
dimensions and social desirability. From an empirical standpoint, Block pointed out that
MMPI variables are related in a meaningful way with extratest behaviors (e.g., Q sort
evaluations; see Block 1965, pp. 97-116).

Finally, Block (1965) derived an MMPI measure of anxiety (based on Welsh's A
factor--see Graham, 1993 for a discussion) that was "desirability free" and found that this
scale correlated with the various MMPI scales essentially the same as when the scale was
not "desirability free."

The SDRS hypothesis was also refuted by Taylor, Carithers, and Coyne (1976).
They suggested that the MMPI acts as a medium through which people can convey "self-
concepts," as opposed to the more narrowly deﬁned construct of social desirability. Self-
concepts were deﬁned as discreet, abstract categories people use to view their behavior
(e.g., moral goodness or badness, controlling or victimized), and people may hold a

number of these concepts to be self relevant. According to this self-concept hypothesis,

35

items on the MMPI are interpreted and answered to reﬂect the individual's self-concept.
This hypothesis was supported in their research.

Apart from issues surrounding what the MMPI K scale actually assesses, another
issue regarding the K scale is a reliable ﬁnding that elevated scores on the K scale are
associated with a higher socioeconomic status (SES) (Dahlstrom, Welsh, & Dahlstrom,
1975; Graham, 1993; Hathaway & McKinley, 1989). Hathaway and McKinley (1989)
suggested that this ﬁnding might be due to people with higher SES being reluctant to
admit to or disclose emotional concerns, doubts, and insecurities as reﬂected in the
MMPI. Conversely, they suggested that lower SES people may believe they have less to
lose by admitting faults or weaknesses. As a related issue, Graham (1993) reported that
on the original MMPI there was a positive relationship between education and K scale
scores. However, on the MMPI-2, Graham stated that the data thus far showed a very
minimal positive relationship between education and K scale scores.

When Graham (1993) summarized the data on the EL, and K scales he stated that
the MMPI has the ability to detect defensiveness (referred to as faking-good), though the
accuracy is not as good as it's ability to detect the fake-bad proﬁle. He further stated that
based on the fake-good proﬁle (i.e., high K low F scores) "One should not infer that the
person is covering up serious psychological problems" (p.53).

In Summary, The K scale has a long and complex history. It is unclear whether or
not the K scale is a measure of defensiveness, psychological health, social desirability,
self-concept, or some other dimension. Further, it is unclear under what circumstances
correcting the clinical scales based on K scores leads to improved clinical discrimination

or a more accurate interpretation. Concerning the K score as a measure of defensiveness,

36

the results are clouded because the vast majority of research conducted with the K scale
used the methodology of asking test-takers to purposely fake good (see Baer et al. 1992).

Thus, although early studies by Meehl and Hathaway (1946/ 195 6) showed
promising results with the K scale as a suppressor variable, the continued use of K as a
corrective score, validity indicator, or measure of defensiveness seems to rely more on
historical precedence than solid research. Dahlstrom, Welsh, and Dahlstrom (1972, pp.
129-130) stated that:

The published research . . . to cross-validate K applications and variations on K

weighting is scanty and inconclusive. Obviously, many agencies within which the

MMPI is used cannot provide ideal criterion information for this kind of cross-

validational work: dependable clinical criteria that are completely independent of

the test-based information or decisions.

It appears that little has changed in the subsequent 25 years of research with the
MMPI.
An MMPI Derived Denial Scale

Apart from the validity and clinical scales, a large number of other scales have
been derived from MMPI items (see Dahlstrom et al. 1975 for a listing), one of which
purportedly measures denial. The Denial scale (Dn) was developed by Little and Fisher
(195 8) from 26 items on the MMPI conversion hysteria (Hy) scale after they made the
observation that the Hy score was often the highest proﬁle point for both psychiatric and
medical patients. Little and Fisher assumed that the elevated scores may have different
meanings for the two groups so they initiated a cluster analysis of the 60 item Hy scale

with data from 60 psychiatric patients and 90 medical patients. The analysis resulted in 9

37

clusters, 5 of which were intercorrelated with each other but not the other 4 clusters--and
vice versa, thus resulting in 2 primary clusters. Eighteen items were left over and they
were added to the cluster with which they were most highly correlated, except 2 items
that were not related to either cluster and thus dropped.

Subsequently, one primary cluster of 32 items was comprised of statements
largely dealing with physiological symptoms that subjects admitted to and was
subsequently called the Admission (Ad) scale. The other primary cluster contained 26
items with statements about poor interpersonal relations, feelings of hostility, suspicion,
and the like, and was named the Denial (Dn) scale. All of the items on the Dn scale
remain on the MMPI-2, though some of the items have been slightly modiﬁed. These
items are reproduced in Appendix B.

Little and Fisher (195 8) reported the correlations between the new scales and
other MMPI scales for a medical patient group and a psychiatric patient group, with 100
participants in each. Based on the 200 cases, the Kuder-Richardson reliability
coefﬁcients were .83 for Ad and .75 for Dn. Ad was most highly correlated with the
MMPI Hypchondriasis scale (r = .89 and .90 for medical and psychiatric patients,
respectively), with which it has an 18 item overlap. Dn was most highly correlated with
the K scale (r = .78 and .88 for medical and psychiatric patients, respectively), with which
it has 9 items in common.

In describing the clinical applications of the Dn scale, Little and Fisher (1958,
p.306) found that "high scores on the Dn scale generally describe the uninsightful, the
anti-intraceptive, the morally virtuous individual; high scores on Dn are also seen with

'muted' or pseudo-normal proﬁles."

38

As mentioned above, the Dn scale is derived from items on the Hy scale.

Trimboli and Kilgore (1983), in an article that supported a psychodynamic approach to
the MMPI (see the introduction of this paper), conceptualized the Hy scale as reﬂecting
the defense mechanism of repression. This conceptualization adds further support to the
notion that the Dn scale is an assessment of denial.

Subsequent research with the Dn scale also adds support to its validity. Watson et
al. (1987) compared the Dn scale with 5 other MMPI derived measures of denial or
repression, as well as with an independently derived measure of repression--the Proj ective
Repression Instrument (PRI). Watson et al. developed the PRI and described it as
consisting of 22 TAT or TAT like pictures that 5 of 6 judges agreed contained sexual or
aggressive themes. Each picture included a multiple choice question that included
choices appropriate to the sexual or aggressive theme of the picture and an additional
choice that ignored the theme--this response, if endorsed, would be considered indicative
of repression. Before being used in the study being described, Watson et a1. tested the
PRI in a pilot study and reported that it is an adequate measure of repression.

Of the 6 MMPI derived measures tested in the Watson et al. (1987) study, the
Little and Fisher (1958) Denial (Dn) scale surfaced as the best measure of repression.

The Dn scale was highly correlated with 2 of the other MMPI measures of denial, and
together these 3 scales were the only ones correlated to the PRI, with the Dn scale being
the most highly correlated.

J orgensen, Schreer, Baskin, and Kolodziej (1992) used the Dn scale in a study that
looked at heart-rate discrepancy scores (HRDS). This study operated from the premise

that defensive operations such as denial inhibit awareness of threatening affective stimuli,

39

however, the stimuli will affect physiological systems such as heart rate (i.e., as denial
increases in response to threatening stimuli, awareness decreases, and heart rate would
increase). The HRDS was derived by taking a participants average report of distress and
anxiety during a rest and stress induced period and subtract those scores from the average
heart rate during the same rest and stress induced period--thus resulting in a discrepancy
score. The results of this study revealed that HRDS was signiﬁcantly correlated with the
Dn scale, as predicted.

Lazarus and Alfert (1964) conducted a study to determine whether priming
subjects (i.e., making commentary statements that may increase or decrease anxiety)
before and\or during a disturbing "subincision" ﬁhn (i.e., the depiction of primitive tribes
performing genital cutting) would affect the subjects' cognitive appraisal and
physiological reaction to the ﬁlm. They predicted that the subjects general tendency to
use denial would impact how they reacted to the ﬁlm. Speciﬁcally, they predicted that
high deniers would verbally report less distress but be more physiologically aroused
(using heart rate and skin conductance monitored every 15 seconds) than those subjects
low in denial, who should show the opposite results. The Dn scale and the MMPI K scale
were among several measures used to assess the trait of denial. Regardless of whether the
subjects were primed for the ﬁlm or not, those that were classiﬁed as high deniers (via Dn
and K) responded as hypothesized--with decreased verbal report of distress and increased
physiological arousal.

McGrath and O'Malley (1986) investigated the Dn scale as well as the Ad (i.e., the
Little & Fisher Admission scale) and the MMPI Hy (conversion hysteria) and Hs

(hypochondriasis) scales on samples of psychiatric, medical, and chronic pain patients.

40

Their results found that the scores on the Dn and Ad scales were positively and negatively
related, respectively, to the presence of denial associated with physical disorders in their
samples.

Two other studies used the Dn scale with results opposite to what has generally
been predicted regarding denial (i.e., increased denial results in decreased self-report of
negative affect and increased physiological reactivity). In the ﬁrst of these studies,
Houston (1973) used the Dn scale to classify his subjects into high or low denial groups.
He subsequently placed the subjects into different experimental groups. In one of these
groups subjects were told that they may receive random shocks during the experiment
(i.e., performing the WAIS Digit Span Task), and another group was told that they would
only be shocked if they performed poorly. There was also a no shock control group and
all subjects were monitored throughout the experiment and were given self-report
inventories of affect before and immediately after the experiment. The results revealed
that those subjects in the high-trait denial group were less physiological aroused and
performed better on the digit-span task. However, it should be noted that Houston (1973)
adjusted each subject's Dn score based on their scores on an anxiety measure. It is
unclear how this adjustment affected the results.

The second study (Mead, 1970) also used the Dn scale to place subjects into low
or high denial groups and looked at the differences between these groups before
undergoing a dental examination. The main intent of the study looked at priming
passages upon the subjects, similar to the Lazarus and Alfert [1974] study reviewed
above. Regardless of which priming passage was used, subjects in the high denial group

self-reported less distress and also had lower EKG's and GSR's. It is of interest to note

41

that subjects in this study had a mean basal EKG heart rate of 92 beats per minute,
compared to 78 in a study that used the "subincision" ﬁlm. One explanation of these
results may be that denial serves a more adaptational role in cases of more extreme threat
(e.g., shock and dentistry) than in less threatening situations (e. g., vicariously
experiencing threat via a ﬁlm). However, it should be noted that the subjects in Mead’s
study were assessed only in anticipation of a dental examination. It would have been
interesting to learn how low and high deniers compared during an actual exam.

Contemporary reviews of the Dn scale are mixed (e.g., Connor, 1986) or avoid
evaluative commentary (e.g., Cramer, 1991). Unfortunately, many of the studies that
have used the Dn scale are older and they do not appear when doing literature reviews
with modern database services, which only access articles back to 1972. For example, a
review by Connor (1986) reported only ﬁnding 1 study that used the Dn scale, whereas
this study review found more than ten.

Levitt (1989) recommended against the use of the Dn scale in clinical
applications. It is unclear if the rejection was made because of a lack of validity studies,
or because of poor or equivocal ﬁndings. However, he positively endorsed the use of two
Harris & Lingoes subscales (Hyl and Hy2, Denial of Social Anxiety and Need for
Affection, respectively) as useful indices of maladjustment or denial. For both of these
scales, all their items (combined N= 18) are contained in the Little and Fisher (195 8) Dn
scale.

In summary, the majority of studies that used the Dn scale found signiﬁcant
results in the predicted directions. Further, the Dn scale has been validated by comparing

it to criterion measures that consisted of more than just other similar self-report measures

42

 

 

(e. g., physiological data). Given its performance to date, the Dn scale seems a
worthwhile measure to investigate in this study.
MMPI-2 Measures of Global Psychological Health

In addition to the validity, clinical, and denial scales mentioned above, the MMPI—
2 contains several various subscales. Relevant to this study, three MMPI-2 scales of
global psychological health will be investigated: the Welsh Anxiety Scale, the Ego
Strength Scale, and the College maladjustment. All three scales are commonly used in
research and clinical settings for an overall assessment of psychological health (versus the
clinical scales, which are more speciﬁc in what they assess; Graham, 1993).

The lVﬂVIPI-2 Welsh Anxiety Scale (A scale; Welsh, 1956), which consists of 39
items, is considered a good index of general psychological distress. In fact the A scale
consists of items based on factor analytic studies, and this ﬁrst-factor cluster has
consistently been labeled as the general distress factor by most researchers (Greene, 1999;
Graham, 1993).

Barron’s (1953) Ego Strength scale (Es scale) is one of the most commonly used
MMPI research scales, and "has been used extensively in both clinical assessment and
research and as an indicator of health, intactness, resilience, psychological and physical
well-being, and as a discriminator between healthy and pathological groups" (Schuldberg,
1992, p. 500). The Es scale was slightly modiﬁed when the MMPI was revised (16 items
were dropped and there were slight word changes on some items), however Schuldberg's
(1992) research on the scale revealed that the modiﬁcations made little difference. The
MMPI-2 Es items are listed in Graham (1993, p. 375). Graham (1993, pp. 137-141)

provides a detailed description of the Es scale, as well as interpretive suggestions (see

43

also Butcher et al., 1990, pp. 69-73; and Levitt, 1989, pp. 74-77, for reviews).

The Mt scale has been used less widely, though a few studies cited in Graham
(1993, pp. 158—159) have found that high Mt scores are associated with therapist rated
distress, and the likelihood of seeking psychological help in college students (though
subsequent research suggested that the scale is more accurate in identifying current
distress). Most of the research with this scale is unpublished; the items are listed in
Graham, p. 376.

Discussion of Marlowe-Crowne-Social-Desirability-Scale

In this study, the 33 item Marlowe-Crowne-Social-Desirability-Scale (MCSD)
will also be used. The MCSD is considered an indirect measure of the need for social
approval, as well as a measure of defensiveness and protection of self-esteem (Crowne &
Marlowe, 1964). The score on this measure will allow this factor to be partialled out of
subsequent correlational analyses in the same manner as defensiveness/denial.

The MCSD has been used to discriminate between a truly low-anxious person
from a repressor (Weinberger, Schwartz, & Davidson, 1979), and is considered to be
unrelated to a person's tendency to answer questionnaires in a socially desirable direction
(Wiesenthal, cited in Weinberger et al.). About half of the items on the MCSD were
chosen because they reﬂected culturally acceptable behavior, though probably untrue for
most individuals, the other half of the items are considered undesirable, but probably true
(Crowne & Marlowe, 1964).

The Assessment of Psychologwealth versus Distress
The assessment of object relations as an indicator of psychological health has

become increasingly accepted as researchers and clinicians have recognized the

44

importance of object relations that occur both intrapsychically and interpersonally (e. g.,
Kissen, 1986; Smith, 1993). There has also been work that related object relations to
physical health and various somatic problems (e. g., Taylor, 1987). Early on, Mayman
(1967) reported that patients' general level of psychopathology could be inferred by their
content responses to the Rorschach as they revealed the quality of underlying object
relationships. More recently, certain changes in object relations have been used as an
indicator of successful change during psychotherapy (e.g., Kavanagh, 1985; Luborsky,
1984)

In the Shedler et al. (1993) study, an experienced clinician (i.e., Mayman) made a
dichotomous rating of subjects as relatively "healthy" or "distressed" based on responses
to the participants' early memory reports (using his Early Memory Test [EMT]).
However, 17 of 58 subjects were unable to be classiﬁed. Though Shedler et al. reported
that interrater agreement regarding the healthy\distressed classiﬁcation on the EMT
ranges from 72 to 80 percent when using very experienced psychoanalytic clinicians, a
review of studies that used the EMT reported poor interrater reliability (Smith, 1993). In
fact, in the Shedler et al. study, they had 2 groups of students score the EMT and
compared these scores to those made by Mayman. The comparison resulted in only a .25
and .37 correlation between each respective group and Mayman's rating.

Another method used in the Shedler et a1. (1993) study to determine whether
subjects were healthy or distressed was a Q-sort. After watching a 60-to-90 minute taped
interview, a clinician performed a Q-sort and based on his response on a single item the
subjects were judged healthy or distressed. Both of these methods of assessment

dichotomized the participants into a health or distressed category, which results in a loss

45

of information and statistical power (see Cohen, 1990, 1992; Van Egeren 1989).
Because neither of these methods appears psychometrically sound, this study will
use the Social Cognition and Object Relations Scale (SCORS; Westen, 1990) to provide a
global rating of psychological health based on responses to several Thematic
Apperception Test cards. This scale was developed from psychoanalytic object relations
theory and social cognition research (Barends, Westen, Leigh, Silbert, & Byers, 1990;
Westen, 1991a). The SCORS assesses 4 independent dimensions of object relations
(Westen, 1991b):
Complexity of Representation of People
This dimension assesses the extent to which the subject clearly differentiates the
perspectives of self and others and recognizes the complexity of the personality
dispositions and subjective experience of the self and others (p. 58).
Affect-Tone of Relationship Paradigms
This dimension assesses the affective coloring of the object world, ranging from
malevolent to benevolent. It determines the extent to which the person expects
relationships to be destructive and threatening or safe and enriching.
Capacity for Emotional Investment in Relationships and Moral Standards
This dimension assesses the extent to which others are treated as ends rather than
means, events are regarded in terms other than need gratiﬁcation, and moral

standards are developed and considered (p. 60).

46

Understanding of Social Causality

This dimension assesses the extent to which attributions about the causes of
people's actions, thoughts, and feelings are logical, accurate, complex, and
psychologically minded (p.60).

Relevant to this study, the SCORS system for the TAT (the system can also be
used with early memories and other data), has been used in research with normal samples,
and the scales have demonstrated their validity with data from interviews, early
memories, and psychotherapy transcripts--all of which also assessed object relations
(Westen, 1991a, 1991b). The Affect-Tone and the Capacity for Emotional Investments
scales on the SCORS were also shown to be correlated with the Sentence Completion
Test for Ego Development (Barends et al., 1990; Westen, 1991b).

Hibbard, Hilsenroth, Hibbard, and Nash (1995) established construct validity of
the SCORS by comparing it to the Concept of the Object on the Rorschach scale (see
Blatt, Brenneis, Schimek, & Glick, 1976). Both of these instruments, as well as others,
were used to assess the same client population from a university clinic. There were
signiﬁcant correlations between relevant subscales on each measure which led the
researchers to conclude that these instruments are assessing what appears to be a
cognitive-structural aspect of human object representation.

This study will use the Affect-Tone (AT) dimension on the SCORS (in
conjunction with Emotional Investment) because it is very similar to the scoring that was
used in the Shedler et a1. (1993) study to assess global psychological health from the early
memories. In fact, Westen (1990) reported that he incorporated the theory and ideas of

Mayman--who scored the early memories in the Shedler study--into his SCORS Affect—

47

 

 

Tone scale. Additionally, after reviewing a description of this study, D. Westen (personal
communication, March 3, 1997) recommended the use of the Affect—Tone scale. He
based this recommendation on his more recent experience with the SCORS that is
currently unpublished. He also reported ﬁnding ample variance on the Affect-Tone
dimension within a normal college sample to allow for discrimination, which is relevant
to this study.

Regarding the validity of Affect-Tone to assess psychopathology, Barends et al.
(1990, p.330) reported that, "in an adult clinical sample, [it] was signiﬁcantly correlated
with the hostility, interpersonal sensitivity, and paranoia subscales of the revised
Symptoms Checklist-90, with [a] Social Adjustment Scale, and with clinician ratings of
interpersonal pathology and nonpsychotic paranoid ideation." A study by Westen, Lohr,
Silk, Gold, & Kerber (1990) revealed that Affect-Tone, as measured on the SCORS,
could distinguish between a non-clinical group, a group with Major Depression, and a
group with Borderline Personality Disorder (from least malevolent to most malevolent,
respectively). Nigg, Lohr, Westen, Gold, and Silk (1992) also found that Affect-Tone, as
assessed from early memories, could distinguish between a Borderline group and a
depressed group. Finally, Fowler, Hilsenroth and Handler (1995), also using Affect-Tone
assessed on early memories, distinguished a university clinic sample from a non-clinical
sample.

In addition to the Affect-Tone scale, the Emotional Investment scale will also be
used to determine if adding EI scores to the AT scale scores increases the overall criterion
validity of the SCORS as evidenced by an increased relationship to the physiological

measures. The use of the El score in this study is considered exploratory, given that there

48

is no direct research that supports its validity as a measure of psychopathology. However,
there is growing evidence that emotional intelligence is an important developmental
personality factor, that may greatly impact several broad areas of a person’s life (e. g.,
Goleman, 1995).

Overall, although the SCORS is a relatively new measure and the majority of
research has been conducted by Westen and his research group, it appears to have
adequate psychometric properties and a comprehensive training manual. Additionally,
two independent reviews of assessment instrumentshnethods of object relations endorsed
the SCORS as a promising measure (Smith, 1993; Stricker & Healey, 1990). The
SCORS AT scale, in combination with the E1 scale, will be used in this study as a
criterion assessment of psychological health which relies on projective techniques and
clinical inference, not on self-report assessment. Finally, the SCORS is one of the few
measures that provides a reliable system of assessment necessary for research.

The Use of Heart Rate and Blood Pressure as Indices of Defense and Distress

The value of physiological measurements, including heart rate (HR) and blood
pressure (BP), to assess psychological states, have long been recognized in stress research
(Katkin, Derrnit, & Wine, 1993). An economic approach is often used to explain how
dealing with stressors over a protracted period will likely result in an increased strain on
the body, which is documented by the physiological measurements (e. g., Wegner &
Pennebaker, 1993; Schonpﬂug, 1986; Selye, 1991).

As discussed in Shedler et a1. (1993), people may report being free of distress,
though clinical impressions and physiological measurements indicated otherwise. It is

ironic that a person's attempt to appear more in control of his or her life by ignoring or

49

denying distressing information may put that person at a higher risk for negative mental
and physical health outcomes. A fair amount of research has been conducted which links
this repressive personality trait to a number of negative outcomes, including increased
anxiety and depression (Larson & Chastain, 1990), as well as elevated blood pressure,
higher rates of cancer, and decreased immune system functioning (Pennebaker, 1985;
Pennebaker & O'Heeron, 1984).

Pennebaker and his associates have moved beyond correlational studies and have
documented that confronting painful affect and distressing information resulted in
decreased autonomic arousal, improved immune system functioning, and a subsequent
decrease in health care visits (Pennebaker & Beal, 1986; Pennebaker, Kiecolt-Glaser, &
Glaser, 1988).

Thus, as demonstrated by Pennebaker's work, as well as Shedler et al.'s (1993)
study, it has become clear that there is often a discrepancy between a person's self-report
of distress vs. the clinical impression, physiological indices, and long-term health
outcomes (see also Bonanno & Singer, 1990; Schwartz, 1990; Weinberger, Schwartz, &
Davidson, 1979; Weinstein, Averill, Opton, & Lazarus, 1968). Because of this
discrepancy, researchers have suggested using physiological measurements to supplement
self-report measures (Temoshok, 1993; Carver, & Matthews, 1989). Given that this study
is a direct attempt to validate a self-report measure, the MMPI-2, as well as building on
Shedler et al.'s work, the use of physiological criterion measures seems imperative.
Further, the choice in using heart rate (HR) and blood pressure (BP) readings will allow
the results of this study to be compared to those of Shedler et a1. (1993), as well as several

other studies that have looked at denial and repression.

50

In addition to HR and BP, mean arterial pressure (MAP) will also be assessed.
MAP has been used in several studies in the stress literature in a manner similar to HR
(e.g., Bruehl et al., 1994; Delistraty, Greene, Carlberg, & Raver, 1992; Haythornthwaite,
Pratley, & Anderson, 1992) and was included here as an adjunct physiological
assessment.

In Sum, physiological measures have been recognized as a valuable addition to
research which allows for an indirect assessment of mental states. The physiological
measures will be used as the second criterion measure to provide an assessment of

psychological health/distress against which the MMPI-2 will be compared.

51

 

 

Method
Participants

Twenty-six male and 31 female volunteers were solicited through the Michigan
State University psychology undergraduate participant pool. The average age of the
participants was 19.76 (SD=2.60), and the range was 18 to 36. Two male subjects were
dropped from the study because one admitted to being drunk during the MMPI-2
administration, the other had an arm so large that the blood pressure cuff would not ﬁt
and BP readings were not possible. The participants received 6 extra-credit points for
approximately 3.0 hours of their time, spread out over the two sessions. Participants with
a self-reported history of cardiovascular problems were asked not to participate in the
study during the ﬁrst testing session.

Instruments
Psychological Measures

Participants completed the ﬁrst 472 items of the MlVfPI-Z (which allowed for the
scoring of all the scales used in this study), as well as the Defense Mechanism Inventory
(DMI) prior to being assessed physiologically. Immediately prior to physiological
monitoring, participants completed the 33 item Marlowe-Crowne-Social-Desirability-
Scale (MCSD).

The MMPI-2 validity and clinical scales were administered and scored in accord
with standard protocol (Hathaway & McKinley, 1989). In addition, the MMPI-2 derived
Denial (Dn) scale (Little & Fisher, 195 8) was scored, as well as the Ego Strength (Es)
scale (Barron, 1953, N=52 items), Kleinmuntz's College Maladjustment scale (cited in

Graham, 1993; Mt, N=41 items), and the Welsh Anxiety scale (Welsh, 1956, A scale, N=

52

 

39). These last three MMPI-2 subscales were used to provide MMPI-2 derived indices of
global health\distress.

The DMI was administered and scored in accordance with the directions
(Ihilevich & Gleser, 1986), which yielded 5 defense cluster scores (described earlier). Of
relevance to this study is the Reversal (REV) score, which provides an assessment of a
person's tendency to use denial\repressive defense strategies. The REV score ranges from
0-to-80 with a higher score representing an increase reliance on denial as a means of
defense. The REV scale will be used as a criterion measure of defensiveness against
which MMPI-2 scales of denial and defensiveness will be compared.

During the physiological monitoring portion of this study, the participants were
administered the following 8 TAT cards: 1, 3BM, 4, 6GP to females, 6BM to males, 7,
10, 13MF, 15 (Westen, 1990 recommends a minimum of 6 cards and 8-to-10 if possible).
The cards were chosen based on those previously used in research with Westen's Social
Cognition and Object—Relations Scale (SCORS, described earlier, e.g., Barends et al.,
1990; Westen et al., 1990), as well as recommendations made by Karon (1968) and
Thompson (1986). The TAT responses were subsequently scored on 2 of the SCORS
scales (i.e., Affect-Tone of Relationship Paradigms, and Emotional Investment) and these
scores were used as an index of global psychological health\distress. For each participant,
the 8 TAT stories were each scored from one through ﬁve, and an average score was
calculated for each participant that was used in the analyses.

To eliminate experimenter bias, the TAT stories were scored by clinical graduate
students and one undergraduate psychology major research assistant, and they achieved

the following interrater reliability scores. On the Affect Tone scale, two graduate-level

53

raters who were previously trained in scoring this instrument achieved very good
interrater reliability (Pearson's correlation=.847, p<.001). On the Emotional Investment
scale, a clinical psychology graduate student and a senior undergraduate psychology
student (who was relatively unfamiliar with obj ect-relations theory) achieved only fair
interrater reliability (Pearson's correlation = .60, p<.001). Interrater reliability was based
on 248 data points, that is 8 TAT scores for each of the ﬁrst 31 participants. Both raters
scored the ﬁrst 31 protocols, then each rater scored 12 participants individually.

Also, during physiological monitoring, participants were given a mental
arithmetic task consisting of items 3-to-14 of the Arithmetic section of the Wechsler
Adult Intelligence Scale-Revised (WAIS-R). A phrase association test was also
administered, which required participants to say the ﬁrst thing that came to their mind
regarding a stimulus phrase. This task consisted of 7 phrases in 5 theme blocks (the
theme blocks included 2 neutral theme blocks, an aggressive theme, a dependency theme,
and a sexual theme). The phrases were adapted from research in this area by Mandler,
Mandler, Kremen, and Sholiton, (1961), and are listed in Appendix C.

The TAT, math and phrase tasks were used in this study to elicit performance
anxiety, and participant’s physiological reactivity was recorded during the administration
of these tasks. The relative increase in heart rate and blood pressure during these tasks is
being used in this study as an indication of increased defensiveness/denial, and
subsequently decreased psychological health (as assessed by the SCORS).

Demographic data was gathered during the second session, prior to physiological
monitoring. Participants provided information regarding their age, gender, education,

occupation, marital status, number, ages and gender of children and siblings (as

54

 

 

applicable), and who lived in their household during childhood. This information is
recommended by Karon (1963) for the interpretation of the TAT.

Information about the occupation and education of the participants' parents was
also gathered to determine the family's socio-economic status (SES). Hollingshead's Two
Factor Index, cited in Myers and Bean (1968), was used for this calculation (the speciﬁc
procedures for this calculation are described in Appendix D). This factor was calculated
because there has been some association between the MMPI K-scale and SES (see the
literature review). However, SES was not signiﬁcantly correlated with the MMPI-2 K or
Dn scale (r=.06 and r=-.01, p>.05, respectively). Given this result, it was not necessary to
partial out the effect of SES when examining the K scale.

Physiological Measures

The measurement of heart rate (HR), systolic blood pressure (SBP), diastolic
blood pressure (DBP) and mean arterial pressure (MAP) was obtained using a Critikon
DINAMAP Vital Signs Monitor, Model 1846. This instrument has built-in programming
which compensates for error variance, such as participant movement during readings. It
also measures actual peripheral pulses vs. electrical signals or contractions from the heart
(electrical signals at the heart do not necessarily produce a peripheral pulse). The
physiological data from this instrument were obtained through the use of pressure cuff
(inﬂated automatically) placed on the participant's non-dominant upper-arm. The
measurements were digitally displayed on the monitor for easy recording, and did not
require any user intervention or calculation. The assessments made with the DINAMAP
were taken in accordance with the operation manual. The physiological readings were

recorded by an assistant and were not visible to the experimenter during the session.

55

 

 

Procedure
Session 1

The ﬁrst session was a group meeting of the participants who volunteered to be in
this study. The participants were seated in a large room with ample surrounding space to
insure privacy while ﬁlling out assessment instruments. The participants were informed
that they were going to ﬁll out personality questionnaires during the session, and that they
would be physiologically monitored while performing various mental tasks during a
second session.

Following the introduction, the participants read and signed an informed consent
form, and they were verbally advised that they could stop their participation at any time
without negative consequences. The participants were also informed, both verbally and.
on the consent form, that their names would not be used, and each participant was
randomly assigned a number that could only be matched to their name by the
experimenter. Participants were asked to exclude themselves from this study if they had a
history of cardiovascular problems.

Next, after brief directions, the MMPI-2 and the DMI were administered. Upon
completion of the instruments, each participant was scheduled for a follow-up
appointment.

Session 2

Upon arrival for the second session, the participants were reminded that they
could terminate testing at any time. To establish rapport, a brief interview was conducted
which allowed for the collection of the demographic data. Next, the Marlowe-Crowne-

Social-Desirability scale was administered. The participant's were then seated in a

56

 

 

 

comfortable chair and the DINAMAP pressure cuff was placed for physiological
monitoring. All participants were then asked to relax for 7 minutes to establish a
physiological baseline. The DINAMAP was triggered to take a reading after 1, 3, 5, and
7 minutes.

Following the relaxation period, participants were administered the TAT, the
mental arithmetic task, and the phrase association task. The order of the 3 tasks was
rotated to provide a counterbalanced design. After each task was administered there was
a 2-minute rest period which concluded with a physiological reading.

TAT Administration

The following directions were read immediately prior to beginning the TAT task:

I'm going to show you a set of [8] pictures, one at a time. I want you to tell me

what's going on, what the characters might be feeling and thinking, what led up to

it, and what the outcome might be. In other words tell me a good story (Karon,

1968,p.89)

These directions were consistent with those made by Westen (1990) in the
SCORS manual. The participants' stories were tape-recorded and subsequently
transcribed and scored using the SCORS. The DINAMAP was triggered to take a reading
as the participant was handed the 2nd, 4th, 6th, and 8th card.

Mental Arithmetic Administration

The mental arithmetic section was introduced with the following directions taken
from Shedler et a1. (1993):

[This] test is a test of mental ability, an IQ test. It is important to try to do well,

because we are going to compare your performance with the performance of

57

 

 

others like you. I'll ask you some questions involving mental arithmetic and you
give the answers. To do well, you must give the correct answer as quickly as you
can. I'll time you with a stopwatch (p. 1120).
The DINAMAP was triggered to take a reading as the 3rd, 6th, 9th, and 12th
problems were read to the participant.
Phrase Association Administration
The phrase association task was introduced with the following directions taken
from Shedler et a1. (1993):
For this [task] I am going to read you some phrases. After each phrase, I want you
to say the ﬁrst thing that comes to mind, as quickly as possible. Give me a
complete sentence or idea, no just a word. Anything you say as a response is ﬁne,
there are no right or wrong answers, but be sure to say the ﬁrst thing that comes to
mind. (p. 1120)
The phrase association task began and ended with a neutral block of 7 phrases.
The DINAMAP was triggered to take a reading at the beginning of each of the 5 blocks.
The responses to the phrases were tape recorded, but were not used in this study.
mg
Physiological Data
The physiological data was used to determine the rate-pressure product (RPP) for
each independent reading taken. The RPP is derived by (HR x SBP)/100 (i.e., heart rate
X systolic blood pressure divided by 100). According to Shedler et al. (1993), the heart is
conceptualized as a pump, and the RPP score captures the heart's rate and force of

strokes. Mean arterial pressure (MAP) is a noncalculated index provided directly by the

58

 

 

blood-pressure monitor. Nineteen physiological readings were taken for each participant.
Eleven of the readings were taken during stressor tasks (4 during the TAT, 4 during the
arithmetic task, and 3 during the non-neutral phrase association task). The last two
readings during the initial rest phrase were averaged together for a baseline reading (the
other readings were taken during rest periods and during the 2 neutral blocks of the
phrase association task).

The RPP was calculated for each observation, and the baseline RPP was
subtracted from each observation during the stressor tasks. Subsequently, RPP change
scores (i.e., from baseline to stressor readings) were aggregated across the 11 observation
points, and resulted in an "average stress change" score which will be used for analyses.
In addition to the average stress change, each participant’s "most extreme change" score--
which consisted of the single observation that yielded the greatest change over the
baseline--will be used for analyses. Given that both of these RPP indices already include
a baseline subtraction, there is no need to statistically control for baseline differences
(however, any signiﬁcant group baseline differences will be noted in text).

Regarding the Mean Arterial Pressure (MAP) indices used in this study, an "average
under stress" score was calculated for each participant by determining their mean MAP
based on the 11 observation points during the stressor tasks. (Note: this index is sensitive
to baseline differences, and the baseline difference must be controlled for statistically if
signiﬁcant group differences exist). The "change from baseline" index was calculated by
subtracting the MAP baseline from each physiological observation point during the
stressor tasks, then the observations during stressor tasks were averaged together for each

participant.

59

 

 

Higher scores on all these indices are interpreted as a stress reactivity score. It was
assumed that the higher this score was, the lower the subject’s psychological health
would be (as assessed by the SCORS scales). This suggests that the individual had
inadequate psychological resources for managing stress when their reactivity score was
high.

Results

Hypothesis 1

It was hypothesized that the MMPI-2 K scale and an MMPI-2 derived measure of
denial (i.e., the Little & Fisher, 195 8, Denial scale) would be meaningfully and
signiﬁcantly correlated (positively) to an independent measure of defensiveness/denial on
the Defense Mechanism Inventory (DMI).

On the DMI, Reversal (REV) was signiﬁcantly correlated with the MMPI-2
Denial (Dn) scale (r=.25, p.<.05), and the correlation of REV with the MMPI-2 K scale
approached signiﬁcance (r=.20, p.=.07). These results provide modest support for the
hypothesis that two of the MMPI-2 scales designed to assess defensiveness/denial are
valid as indicated by their relationship with an independent assessment of these defenses.
Additionally, it should be noted that K and Dn were correlated more with the REV scale
than any of the other four DMI defense cluster scales. (Note: All correlations reported are
one-tailed, and the alpha level was set at .05).
Findings regarding the Social Cognition and Object Relations Scale and the Physiological
Reactivity Scales

In contrast to study expectations, as shown in Table 1, the Affect Tone (AT) scale,

and the AT scale in combination with the Emotional Investment (E1) scale, were

60

 

 

nonsigniﬁcantly correlated with the four physiological indices that were described earlier
(results for the E1 scale alone are also shown). The AT and EI scales, which are part of
the Social Cognition and Object Relations Scale (SCORS), were used in this study to
provide a global index of psychological health/distress. It was anticipated that there
would be a meaningful correlation between these indices and physiological reactivity
scores, whereby those participants with more positive affective tone in their relationships
(i.e., displayed in their TAT stories) would display less physiological reactivity under
stressed conditions.

This ﬁnding will make it necessary to do separate analyses by comparing the
MMPI-2 assessments of psychological health independently to both the physiological
reactivity scores, as well as the SCORS assessment of psychological health. For reasons
that will become clear later (discussed below), of the two indices of psychological health-
-the SCORS assessment and the physiological reactivity scores--the physiological indices
are considered to be more valid, and will be used as the main study criterion measure of
psychological health (i.e., the capacity to cope with stress).

Hypothesis 2

Based on the ﬁnding of the lack of speciﬁcity of self-report measures of health in
the Shedler et al. (1993) study, it was predicted that the MMPI-2 would not provide an
accurate assessment of global psychological health\distress for those participants that
appear healthy. This lack of speciﬁcity should be reﬂected in a lowered overall
correlation with two indices of psychological health\distress (i.e., a clinical impression
and physiological measures of heart rate and blood pressure), unless defensive denial is

taken into account. This hypothesis would be conﬁrmed if the effect of partialling out

61

denial--as assessed by the MMPI-2 K scale or an MMPI-2 derived scale of denial—
signiﬁcantly increased the correlation between the MMPI-2 assessments of psychological
health\distress and the criterion measures of physiological reactivity, as well as increased
the correlation with scores on the Westen SCORS scales.

MMPI-2 and Physiological Reactivity Scores

As shown in Table 2, the three MMPI-2 measures of global psychological
health/distress were not signiﬁcantly correlated with the physiological reactivity indices
used in this study as criterion measures of psychological health. However, it should be
noted that the College Maladjustment Scale approached signiﬁcance with the RPP most
extreme change index (r=.20, p.=.07), and the Welsh Anxiety Scale approached
signiﬁcance with the MAP change from baseline index (r=.20, p.=.08). (Note: These
correlations are decreased by error variance due to over-estimating and under-estimating
the psychological health of participants. Obviously, a decrease in errors due to over-
estimating psychological health will be reﬂected in a decrease in total error variance,
which will increase the correlation. Therefore, the Pearson Product Moment Correlation
provides an “efﬁcient” [in the statistical sense] statistical index of increases or decreases
in these types of errors).

Table 3 shows the correlations between the three MMPI-2 measures of global
psychological health and the physiological reactivity scores while statistically controlling
for denial by partialling out the MMPI-2 Dn scale of the correlational analyses. The
results from these analyses generally supported the hypothesis that the correlation
between the MMPI-2 assessments of psychological health and the physiological indices

would signiﬁcantly increase when denial was taken into account. More speciﬁcally, the

62

 

MMPI-2 Anxiety and College Maladjustment scales were signiﬁcantly and positively
correlated (while denial was partialled out) with several of the physiological indices.

Although not shown, the above set of partial correlational analyses between the
MMPI-2 psychological health assessments and the physiological reactivity scores were
also run while independently partialling out the effects of the MMPI-2 K scale and L
scale, the Marlowe-Crown scale, and the DMI REV scale--all of which are considered
measures of defensiveness/denial. The most consistent and signiﬁcant results that
supported the study hypotheses were found with the MMPI-2 Denial scale, which is most
consistent with the study hypothesis. The results from partialling out the various other
denial scales generally followed the results above, but to a lesser degree.
MMPI-2 and Sociﬂognition and Object Relations Scales

As predicted, the three MMPI-2 measures of global psychological health/distress
(i.e., the Ego Strength Scale, College Maladjustment Scale, and Welsh Anxiety Scale)
were not signiﬁcantly correlated with the SCORS Affect Tone (AT) or AT+EI
(Emotional Investment) scores (see Table 4). However, even when denial was controlled
for, the partial correlations (controlling for denial with the MMPI-2 Dn scale, as well as
the K and L scales) between the MMPI-2 and the SCORS AT and AT+EI scales remained
nonsigniﬁcant (see Table 5) except for the Ego Strength scale, which was signiﬁcantly
correlated with the AT+EI scale in the opposite direction of the study prediction (recall

that lower ES scores indicate increased distress).

63

 

Hypothesis 3

It was hypothesized that as the measure of denial increased, the physiological
reactivity scores would increase, and the clinical impression of health based on the
SCORS would decrease.
MMPI-2 Denial and Physiological Reactivity Scores

As shown in Table 2, there were no signiﬁcant correlations between the MMPI-2
measures of defensiveness and denial (i.e., the K and Dn scales) and the physiological
reactivity indices. These ﬁndings do not conﬁrm this hypothesis.
MMPI-2 Denial and SCORS

Shown in Table 4, the MMPI-2 measures of defensiveness and denial were
signiﬁcantly and positively correlated with the AT scale, and the Dn scale was
signiﬁcantly and positively correlated with the AT and EI scale combination. These
ﬁndings are in fact in the opposite direction of the hypothesis. This ﬁnding indicates that
as defensiveness and denial increase (as assessed on the MMPI-2), the participants look
more psychologically healthy on the SCORS scales, a ﬁnding which may help explain the
unexpected results in this study with respect to the SCORS scales.
Summgy

The results supported the validity of the MMPI-2 Denial scale (and to a lesser
extent the K-scale) as a measure of defensiveness/denial. This ﬁnding conﬁrmed the ﬁrst
hypothesis. The second hypothesis was conﬁrmed with respect to the physiological data,
but not the SCORS data. The third hypothesis was not conﬁrmed with the physiological

data, and there were actually opposite ﬁnding with respect to the SCORS data.

64

 

 

Exploratory Analyses of the Study Measures of Psychological Health

Given the unexpected ﬁnding that there was no relationship between the SCORS
assessment of psychological health and the physiological reactivity scores, exploratory
analyses were conducted to better understand the relationship between these variables.

A mean split was performed on the participants' Affect Tone (AT) scores, and the
average physiological reactivity scores were calculated for each group. Individual scores
on the AT scale ranged from 2.19 to 3.63, and the overall mean score was 2.79 (SD=.31)
which was the dividing point. As would be expected given the correlational ﬁndings, as
shown in Table 6, there was a nonsigniﬁcant difference (a slight trend in the wrong
direction) between the average physiological reactivity score (i.e., the RPP average stress
change score) for the distressed group (those with AT scores below 2.79; n=3 0) versus
the healthy group (those with AT scores above 2.79; n=25), t(41)=-.49, p=.63, two-tailed.

The two groups above (i.e., high and low AT scores) were also analyzed on their
most extreme physiological reactivity score achieved during the various stressor tasks
(see the Scoring section for more information). This analysis also revealed a
nonsigniﬁcant difference between the two groups t(53)=.10, p=.92, two-tailed (see Table
6).

The mean arterial pressure (MAP) was also calculated for the high and low AT
groups. To begin with, the low AT group had an overall lower baseline MAP mean score
M=83.05 (SD=6.08) than the high AT group M=87.60 (SD=9.01), t(53)=-2.23, p=.03,
two-tailed. Because there was a signiﬁcant baseline difference, an ANOVA was
performed to determine the main effect of the AT group difference while using the

baseline as a covariate. This analysis revealed a nonsigniﬁcant AT group effect F(2,

65

54):.008, MSE=.306, p=.93. Finally, there was a nonsigniﬁcant MAP change from
baseline score between the healthy and distressed AT groups (p=.44, two-tailed; see Table
6).

To widen the gap between the participants that looked healthy versus distressed
on the Affect Tone scale, the participant's with AT scores around the mean were removed
and analyzed separately, and were viewed as "undetermined" regarding their
psychological health status. As shown in Table 7, there was still no signiﬁcant difference
between the high AT group and the low AT group on any of the physiological indices. An
interesting ﬁnding was that the undetermined group (i.e., those with AT scores around the
study mean) had signiﬁcantly lower reactivity scores on several indices than both the high
and low AT groups.

To determine if the ﬁndings above were due to outliers, the participants individual
AT scores and psychophysiological data were examined, and the results were clearly not
due to a small number of outliers. For example, of the 14 participants in the AT healthy
group, 10 participants RPP average stress change scores exceeded the study mean on this
index--and 4 of these 10 were approximately 2 or more standard deviations above the
mean. These ﬁndings are opposite than what was anticipated--that is, they reveal that the
participants with higher AT scores have greater physiological reactivity while stressed
than the participants with lower AT scores.

As mentioned in the Method section, another index from the SCORS TAT
instrument would be used in conjunction with the AT scores to determine if the combined
scores added criterion validity. The Emotional Investment (E1) scores were averaged

with the AT scores into a combined scale and analyses were performed identically to

66

those conducted above with the AT scores alone. The results were nearly identical. That
is, there were no signiﬁcant results. Thus, adding E1 to the AT scores did not increase the
criterion validity of these scales when compared with the physiological indices.

Thus, all the analyses failed to show any relationship between the SCORS scales
and the physiological reactivity reactions to stress (with the exception of the
“undetermined” group).

Unstructured Impressions of Global Psychological Health

Given the unanticipated ﬁnding regarding the lack of relationship between the
SCORS assessment of psychological health and the physiological reactivity scores, a
decision was made to rate global psychological health/distress based on the participant's
responses to the TAT cards using a method similar to the one used in the Shedler et al.
(1993) study. That is, in addition to using the Social Cognition and Object Relations
Scale to assess the TAT responses, an unstructured clinical assessment was based on the
affective tone of the responses. In the Shedler et a1. study, psychological health was
based on clinical impressions of early memories gathered from study participants.
Mayman (one of the study authors) then assessed the memories:

[By attending] . . . to the qualitative factors such as how the self was represented,

how the interpersonal world was represented, the affective tone of the material,

whether the memories were narratively coherent or contained inner contradictions

(suggesting omissions and distortions), and so on. (p. 1121)

In this study, after becoming familiar with Mayman's work (e.g., Mayman, 1967)
a similar technique was applied to the TAT responses by the experimenter (Heim). While

making the assessments, the experimenter was blind to all other data, and it should be

67

 

 

noted that he did not participate in scoring the TAT responses using the SCORS. It
should be further noted that the assessments made in this study were not cross-checked by
an independent rater, and no attempt was made to validate the experimenter impressions.
The assessment was simply done as an additional exploratory process to better understand
the data.

An example of stories that were rated healthy versus unhealthy are contained in
Appendix E, with an explanation of how they were rated. Basically, each of the 8 TAT
responses from each participant were read and judged as healthy, distressed, or neutral
(i.e., unable to draw a conclusion). Based on how the 8 stories were rated, an overall
assessment was made which classiﬁed the participant as psychologically healthy,
distressed, or undetermined. Of the 55 participants, 10 were judged as healthy, 26 were
judged as distressed, and 19 were undetermined. This breakdown of global assessments
was similar to the Shedler et al. (1993) study. In that study, of 58 participants, 12 were
judged as healthy, 29 were judged as distressed, and 17 were left unclassiﬁed.

After the classiﬁcations were made, the different groups were compared on their
RPP average stress change scores during the stressed periods. The group that was judged
healthy had a mean reactivity score of M=10.51 (SD=9.64), the group judged distressed
had a mean score of M=14.36 (SD=13.09), and the undetermined group's score was
M=14.78 (SD=10.69). A t-test between the groups judged healthy versus distressed was
non-signiﬁcant: t(34)=-.84, p=.41. However, in contrast to the results based on the
SCORS AT assessment, the current ﬁndings displayed a trend in the expected direction,
though not statistically signiﬁcant (i.e., less physiological reactivity while under stressed

conditions in the group judged healthy than the group judged distressed).

68

Summary of Exploratog; Analyses Between the SCORS and the Physiological Indices

In summary, there were nonsigniﬁcant differences in physiological reactivity
between "healthy" versus "distressed" groups based on AT scores. Even when more
extreme groups were examined by removing those participants with AT scores around the
mean, there were still no signiﬁcant group differences between the AT assessed healthy
versus distressed groups. One interesting ﬁnding was that the undetermined group (those
with AT scores around the mean) had signiﬁcantly lower reactivity scores than the both
the low and high AT scoring groups. When the TAT stories were scored using an
unstructured clinical impression by the experimenter, the results, though nonsigniﬁcant,
showed a trend in the anticipated direction (i.e., those judged healthy had less
physiological reactivity). Possible explanations for these results will be presented in the
Discussion section below.
Exploratory Analyses Based on Group Classiﬁcations by MMPI-2 Scores in Conjunction
with Affect Tone Scores

Although there was not the anticipated relationship between the two indices of
psychological health used in this study (i.e., the SCORS and physiological indices),
additional analyses (outlined below) were conducted in the same manner as what was
described in the Shedler et al. (1993) study. Given the lack of relationship between the
two study indices of psychological health, these analyses were performed for illustrative
purposes only, and the results were not expected to be signiﬁcant--and in fact none were
signiﬁcant.

The following analyses focused on the MMPI-2 Welsh Anxiety Scale (A scale),

which is considered a good index of general psychological distress, and showed the most

69

 

 

 

signiﬁcant results in this study in the preceding analyses. These analyses were originally
designed to combine the assessment of psychological health based on the MMPI-2 A
scale in combination with the SCORS AT assessment to try and predict the
participants’physiological reactivity scores while under stressed conditions.

Participants were divided into 4 groups based on their SCORS Affect Tone scores
in combination with their MMPI-2 Welsh Anxiety scale (A scale) scores. Three of these
groups were examined as relevant to the hypotheses. On the A scale the study mean was
55, but 58 was used as a cut score to distinguish healthy versus not healthy groups based
on a natural break in the data (scores above 60 are generally interpreted as falling in the
distressed range, e. g., Greene, 1990). If the participant looked healthy on both measures
(MMPI-2 A scale<58 and SCORS AT>2.85), they were considered "truly healthy". If the
participants looked healthy on the MMPI-2 A scale (A<5 8), but distressed on the AT
score (AT<=2.85) they were considered to have "illusory health" as coined in the Shedler
et a1. (1993) study (i.e., the person looks healthy on a self-report measure, but distressed
on a more sensitive, clinically judged impression). Finally, the third "truly distressed"
group was comprised of those participants that appeared distressed on both the MMPI-2
A scale and the AT scale (A>=58 and AT<=2.85). The RPP average stress change was
compared between the 3 groups. (Note: the terms “truly healthy”, “illusory healthy”, and
“truly distressed” are based on the combination of MMPI-2 A scale scores and the
SCORS AT scores. These terms will be used in this section, but are not meant to convey
the participants’ actual real-world psychological health).

Based on ﬁndings from the Shedler et a1. (1993) study, it was originally predicted

(before the realization that there was no relationship between the SCORS and

70

physiological measures) that there would be signiﬁcantly greater average physiological
reactivity during stressor periods for the "illusion" group (because they are more likely to
be distressed) than the “truly healthy” group. However, as shown in Table 8, there were
no signiﬁcant differences in the mean physiological reactivity between any of the three
groups.

Additional analyses were performed with the other 3 physiological indices (see
discussion in the Scoring section for an explanation of these scores) that examined the
differences between the healthy, illusion, and distressed groups. The results for these
analyses are also shown in Table 8, and again no signiﬁcant group differences were
revealed. Finally, analysis were conducted that used different SCORS Affect Tone cut
points to distinguish healthy versus distressed participants (see Table 9), and further, the
combination of the SCORS AT and E1 scores were used to distinguish healthy versus
distressed groups (see Table 10).

Overall the results from these analysis regarding the combination assessment of
psychological health based on both the MMPI-2 Anxiety scale and the SCORS AT scale
can be summed up as follows. There were no signiﬁcant results. Further, the trend of the
data (based on nonsigniﬁcant results) were often in the opposite direction of what was
anticipated (i.e., the group that scored healthy on the combination assessment of the
MMPI-2 A scale and AT scale showed a trend toward greater physiological reactivity
under stressed conditions than the group that was assessed distressed on those two
measures. Finally, all of these analyses were run with other MMPI-2 measures of
psychological health/distress (i.e., the Ego Strength scale, College Maladjustment scale,

and the average of 9 MMPI-2 clinical scales)--similar nonsigniﬁcant results were found.

71

 

Discussion

The intent of this study was to extend and generalize the ﬁndings of the Shedler et
a1. (1993) Illusion of Mental Health study by examining MMPI-2 global measures of
psychological health/distress. This would be accomplished by comparing various MMPI-
2 measures of psychological health against both a clinical impression of psychological
health (based on TAT stories scored by Westen’s Social Cognition and Object Relations
Scale), and a criterion measure of physiological reactivity scores (heart rate and blood
pressure reactivity scores under stressed conditions). The validity of the MMPI-2
measures are in question because they rely on self-report, and, as shown in the Shedler et
al. (1993), such measures are susceptible to the effects of denial that may moderate their
results—particularly when the results indicate psychological health. This study was
designed to extend the results of the Shedler study, by actually testing the MMPI-2, using
methodology very similar to that used in the Shedler et a1. study. However, the present
study contained some important differences (i.e., the use of the SCORS to provide a
clinical impression of psychological health, and the MMPI-2 was used as the self-report
measure) which will be explored below. The problems that were found with the SCORS
measure limited the ability of the present study to examine the validity of the MMPI-2.
However, some results were signiﬁcant and are worth noting.
Major Findings

This study provided some support for the validity of two MMPI-2 measures of
defensiveness and denial. The Little and Fisher (195 8) Denial scale (Dn) showed better
psychometric properties than the K scale as evidenced by correlations with an

independent measure of denial on the Defense Mechanism Inventory Reversal (REV)

72

scale. Further, when these measures were partialled out of subsequent correlational
analyses, the results revealed the anticipated increase in the correlation between MMPI-2
global measures of psychological health (i.e., the Anxiety and College Maladjustment
scales) and the physiological reactivity scores while under stress--which used as an index
of psychological health in this study.

The current results also suggest that the MMPI-2 Anxiety scale (A scale), and to a
lesser extent the College Maladjustment scale (MT scale), appear susceptible to the
effects of denial. Thus their assessment of psychological health may indeed include an
“illusory”component, whereby a truly distressed person may appear healthier than he or
she really is on these MMPI-2 scales. However, if denial is taken account (i.e.,
examining the person's score on the MMPI-2 Dn scale), a more accurate assessment of
psychological health can be obtained by realizing that the person's scores on the MMPI-2
Anxiety and Maladjustment scales may be underestimated.

This is the principle used in adjusting several of the MMPI-2 clinical scale scores
based on K scale scores. However, according to Graham (1993), few if any studies on the
MMPI-2 have been conducted to determine if the K corrected scores increase the validity
of the various clinical scales. The results of this study suggest that when using the
Anxiety scale or the Maladjustment scale, the assessment of global psychological
health/distress can be improved if the Dn scale scores are known and taken into account.
(Note: the clinical scales were not used in this study because they are more speciﬁc in
their assessment of distress, whereas the scales used in this study provide a more global
assessment of psychological health/distress. As a result, no conclusion can be made

regarding the validity of using the K scale to correct the clinical scale scores).

73

Findings Regarding the Social Cognition and Object Relations Scale

As mentioned above, the results from this study are limited in their ability to
extend the Shedler et a1. (1993) ﬁndings because the two independent measures of
psychological health against which the MMPI-2 was to be compared were not correlated
with each other. Possible explanations will be explored.

As reported in the results, there were almost no signiﬁcant differences in rate
pressure product (RPP) changes while under stressed conditions between groups that
were assessed “healthy” versus “distressed” by the SCORS AT measure. Even using
correlational analyses, which yields more statistical power than a comparison of means,
none of the physiological indicators showed the expected relationship to the SCORS AT
(or AT+EI) assessment of global health/distress.

A problematic factor of this study seems to have been the use of Westen’s (1990)
Social Cognition and Object Relations Scale (SCORS) instrument to develop the clinical
impression of psychological health used in this study. Whereas Shedler et a1. (1993) used
an unstructured clinical impression of early memory data to assess global psychological
health (the Early Memory Test, scored by Mayman), the present study used TAT
responses that were scored by graduate student (and one undergraduate) using the
SCORS.

In both this study and the Shedler et a1. (1993) study, both sources of data (i.e.,
early memories and TAT stories) were scored based on the affective quality of the
material generated by the participant. In fact, Westen (1990) acknowledged incorporating
the theory and rational described by Mayman (who scored the early memories in the

Shedler study) into his SCORS instrument. However, although the data were scored

74

 

similarly, they were generated from very different sources.

That is, the early memory data is self-generated, and the data in this study was
generated in response to visual cues provided by the TAT. Although both sources of data
likely contain important information about the participant providing the response, it may
be that the self-generated early memories offer more unadulterated insight into the
person's psychological functioning. That is, in terms of structure, the TAT provides more
structure than early memories, and as structure increases, the projective element
decreases.

This difference in the source of data allows for speculative explanations for the
nonconvergent ﬁndings between this study and the Shedler et al. (1993) study. For
example, in this study some of the pictures on the TAT cards may set the stage for
responses that contain affective material which represents discord between the story
characters (i.e., card pull). This content would subsequently receive a lower score on the
SCORS AT scale, indicating less psychological health. Even Westen (1991b) discussed
that there is an element of card pull which sets up the story teller to give relatively
ordinary malevolent responses.

Participants that ignore the card pull and give responses that "deny" the apparent
content (i.e., their response contains a great deal of unity and harmony in what looks like
a more tragic scene) will receive higher AT scores on the SCORS. In other words, those
participants who engage in denial--and may be pollyannish--receive the highest AT scores
and “appear” psychologically healthy.

Following this line of reasoning, high scorers on the AT scale may actually

represent a selection of participants that use psychological defense of denial, and who

75

may be less psychologically healthy than participants who had moderate scores on the AT
scale because these participants may have more realistically dealt with the content of the
TAT cards.

Evidence for this proposition is shown in Table 11. All of the measures in this
study that are designed to assess defensiveness (MMPI-2 K scale), denial (MMPI-2 Dn
and the Defense Mechanism Inventory REV scale), and naive attempts to present oneself
in a favorable light (MMPI-2 L scale and the Marlowe Crowne scale) are signiﬁcantly
and positively correlated with the AT scale. That is, as a person looks psychologically
healthier on the AT scale, they appear to engage in increased defensiveness/denial--
according to a wide range of different measures.

Further evidence for this proposition was provided by a post-hoc correlational
analysis that was run which did not include the 14 participants who scored “healthy” on
the SCORS AT measure. This analysis revealed a moderate negative correlation between
the AT measure and the average physiological reactivity while under stressed conditions
(r[4l]=-.29, p=.03). This is the ﬁnding that was expected for the entire study population
(i.e., as global health increases as assessed by the SCORS AT scale, physiological distress
decreases).

The physiological data also seems to support the proposition that the AT scale is
susceptible to the inﬂuence of denial. A review of Table 7 shows that the "undetermined
group" (i.e., those with moderate scores on the AT index) had signiﬁcantly less
physiological reactivity while under stressed conditions than the “healthy” group (i.e.,

those participants with high AT scores).

76

 

Other data also seem to conﬁrm that as AT scores increase beyond a certain point,
some of the participants (i.e., those who rely on denial) at the high end of the scale may
actually be less psychologically healthy (as evidenced by increased physiological indices),
and thus washing out any correlational signiﬁcance between the SCORS AT scale and the
physiological reactivity scores. A review of Tables 6 and 7 shows that the variability of
the AT “healthy” group is overall greater than that for the “distressed” group and the
“undetermined” group (in Table 6, a Levene’s Test for Equality of Variances revealed
that there is signiﬁcant difference between the “healthy” and “distressed” groups’
standard deviation means for the RPP average stress change: F=4.94, p.=.03). This
ﬁnding suggests that the those participants who score at the higher end of the AT scale
appear to represent a cross-section of people—some who may be healthy, and some likely
unhealthy (based on physiological reactivity scores).

Only one previous study could be located that compared the SCORS to the
MMPI-2. In this study, the SCORS was signiﬁcantly and positively correlated with the
psychotic triad on the MMPI-2, especially the Schizophrenia scale. This ﬁnding was in
the opposite direction of the study's expectations (Hibbard et al., 1995). This result lends
additional support to the notion that the AT scale seems to be measuring a different
construct than psychological health, or that it may be easily manipulated into providing an
illusory assessment.

There was one line of evidence that does not support the above proposition.

When denial was controlled for statistically (i.e., by partialling out the various measures
of defensiveness/denial used in this study) there were still no signiﬁcant correlations

between the AT scale scores and the various physiological measures. This ﬁnding

77

suggests that there may be other constructs besides (or in addition to) denial that
moderate the AT scale, or the AT scale may measure a construct not related to
psychological health--at least as compared to the physiological measures.
In sum, it appears that the same dilemma can be applied to the SCORS AT

measure as was applied to self-report measures by Shedler et al. (1993). That is, as a
mental health scale, the AT scale may be valid when it indicates distress, but the results
are ambiguous when it indicates health because the person may be truly healthy, or they
may be engaging in psychological defense of denial--which is presumed unhealthy. It is
ironic that the criterion measure--which was chosen to provide a clinical impression of
psychological health free of defensive distortion--seems as susceptible to psychological
defense as self-report measures.

It should be stated that the SCORS AT index was not intended by Westen (1990)
to be used strictly as a measure of psychological distress. That is, it may be more a
measure of maturity or some other personality factor not necessarily equated with
distress, even though the research reviewed earlier indicated that it does assess distress.
Further, Westen reported that the SCORS represents a continuum of development of

object relations which does not necessarily parallel a continuum of pathology (see the

review by Smith, 1993).

Apart from the reasons mentioned above, other general factors might account for
the discrepant ﬁndings between this study and the Shedler et al. (1993) study regarding
the nonsigniﬁcant relationship between the clinical impression of psychological health
and the physiological indices. For example, the participants from either study may have

been a non-representative sample of the population. In fact, the participants in this study

78

 

 

 

'ere drawn from the MSU psychology subject pool, and a recent article in The State
lews (“Undergrads Used”, 1999, September)--the campus newspaper at MSU—raises
oncerns about this study’s participants. In the paper, an undergraduate student wrote an
pinion article stating that he does not like having to participate in university research
which is required in several courses at MSU), and as a result, his intention is to give
ivalid responses to researchers, and generally be as non-cooperative as possible. He
lluded to the fact that several other students beside himself also do not take their roles as
esearch participants seriously.
Conclusion

Although this study was limited in its ability to extend the Shedler et al. (1993)
tudy, some ﬁndings deserve mention, and further discussion of the Shedler et al. study is
varranted. The Shedler study reported that self-report mental health measures may be
nvalid indicators of psychological distress because psychological defense (i.e., denial)
nay lead to an illusion of mental health, whereby a truly distressed person that denies
heir distress (either consciously or unconsciously) is able to appear healthy on a self-
eport measure. They suggested that their results apply to the MMPI-2, even though they
lid not use this measure in their study. Based on the results of the present study, as well
LS a careful examination of the Shedler et al. study, applying their ﬁndings to the MMPI-2
tppears premature or unwarranted at this time.

In the Shedler et al. (1993) study, the Eysenck Neuroticism (N) scores were used
ndependent of the built in validity scale (the L scale). Eysenck (1994), in a rejoinder to
he Shedler et al. study, reported that "it is not permissible to interpret N scores

ndependently of L scores" (p. 972), which is what Shedler et al. did in their study.

79

 

 

 

 

In fact, Shedler et al. (1994) did acknowledge that mental health scores should not
be interpreted in isolation (i.e., without checking validity scales), however, they went on
to discuss that researchers frequently do use mental health scale scores at face value,
which perpetuates the idea that research ﬁndings in psychology should not be trusted.
However, it seems that the point being overlooked by Shedler et al. is that the fault with
"illusory" ﬁndings may be more with the researchers than with the instruments.

Admittedly, there are many mental health scales that are straight-forward, face
valid instruments such as the Beck Depression Inventory (Beck, Ward, Mendelson,
Mock, & Erbaugh, 1961) that make no attempt to check for psychological defense (i.e.,

include validity scales), and these scales are likely to be susceptible to illusory ﬁndings as

reported by Shedler et al. (1993).

Compared to the Eysenck Neuroticism scale, the MMPI-2 contains several more
validity scales, some of which are quite sophisticated, as well as a much larger research
base. It would be unwise, if not unethical, to interpret an MMPI—2 proﬁle without taking
into account the overall validity of the test results. Even when using MMPI-2 subscales,
as was done in this study, the validity indices should not be ignored. In fact, there was
modest support in this study that one of the most widely used MMPI-2 subscales, the
Welsh Anxiety scale, appears to have increased validity if denial is taken into account.

Now that many of the more popular psychological instruments are computer
scored, it is easier to view a wide range of validity scales and subscales that present a
more sophisticated view of the test-taker. For example, Green's computer report for the
MMPI-2 includes a number of validity indicators beyond L, F and K, (e. g., % True,

%False, VRIN, TRIN, F -Fb) and he strongly endorses the regular use of these scales

8O

efore even looking at the clinical scales or subscales (Green, 1999).

It should be stated that it takes a great deal of training and experience in using the
IMPI-2 to be able to accurately interpret the test results. Experience seems to be a key
rctor in making an accurate assessment of psychopathology--regardless of the data used
> make the assessment (i.e., structured or unstructured tests, clinical interviews, etc.).
or example, in the Shedler et al. (1993) study, Mayman, an experienced clinician, was
:portedly better at rating study participants as healthy versus distressed than student
1dges.

One of the main conclusions in the Shedler et al. (1993) study was that "clinical
rdgments provided information about mental health that was, apparently, not available
rom 'obj ective' mental health scales" (p. 1129). I agree that a careful, clinical judgment
y an experienced clinician can provide invaluable information about a person, although
iere was no empirical support for that in the present study.

However, I also believe that a highly trained, experienced person using the
/IMPI-2 could provide a reasonably accurate and valid assessment of the test-taker. In
ases where the test results are not clear--for example there is an indication that the
rerson may be underreporting-it would only be prudent to acknowledge that fact, and not
lraw conclusions about the person's psychological health without other sources of data.
{ecall that in the Shedler et al. (1993) study that nearly one-third of their participants
vere left unclassiﬁed regarding their psychological status because the data were

nconclusive.

8’1

 

 

Finally, an issue apart from the MMPI-2 deserves mention. In the Shedler et al.
.993) study, they reported that of the 41 subjects they were able to classify 29 (71%) of
rem fell in the distressed category. Being as conservative as possible, if the 17
nclassiﬁed subjects were all placed in the healthy group, that would still leave 50
ercent of their college sample as distressed.

As a comparison to Shedler et al.'s (1993) rate of distress, one of the most
omprehensive studies to date of psychiatric disorders in the United States, the
ypidemiological Catchment Area study (Robins, Locke, & Regier, 1991), reported a 20%
nnual prevalence rate for all diagnosable disorders.

In response to this criticism of their study (e. g., Joiner, 1994) Shedler et al. (1994)
eplied that they may have a more differentiated and specialized view of mental health
ran that of most mental health researchers. This specialized view does not rely on
ymptomatology, and they stated that they perceive mental health as the capacity to form
atimate relationships, experience a range of affect, to be creative, and the like.

How distress is deﬁned obviously affects how many people will be classiﬁed in
hat category. Given that Shedler et al. (1993, 1994) have more stringent criteria to be
lassiﬁed as healthy, this may help explain why the self-report measures they studied
eemed inadequate. That is, most self-report measures were not designed to assess the
ype of distress they outlined.

In sum, the conclusion of the Shedler et al. (1993) study that self-report measures
>f psychological health may result in illusory health may have been overstated based on
heir data--at least regarding the MMPI-2. Although self-report scales may report illusory

realth at ﬁrst glance, careful attention to validity scales, an understanding of defensive

82

rocesses, and signiﬁcant experience with the test, should help ensure a reasonably
ccurate assessment of distress (or at least an acknowledgment of being unsure).
[owever, their research did reveal that the process of denial in a test-taker can
igniﬁcantly impact their scores, and this point can not be summarily dismissed. The
roblem of denial was also present in the SCORS measure used in this study. Given the
rnited ability of the present study to extend and generalize the ﬁndings of Shedler et al.,
dditional research to check the validity of self-report measures, including the MMPI-2,
eems imperative. Finding valid, reliable criterion measures to conduct these studies is

.ey, and will likely make the endeavor difﬁcult.

83

 

REFERENCES

Anastasi, A. (1988). Psychological Testing (6th ed.). New York: Macmillan
ublishing Company.

Baer, R.A., Wetter, M.W., & Berry, D.T.R. (1992). Detection of underreporting
fpsychopathology on the MMPI: A meta-analysis. Clinical Psychology Review, 12,
09-525.

Barends, A., Westen, D., Leigh, J ., Silbert, D., & Byers, S. (1990). Assessing
ffect-tone of relationship paradigms from TAT and interview data. Psychological
,ssessment: A Journal of Consulting and Clinical Psychology, 2, 329-332.

Barron, F. (1953). An ego-strength scale which predicts response to
sychotherapy. Journal of Consulting Psychology, IL 327 -333.

Beck, A.T., Ward, C. H., Mendelson, M., Mock, J ., & Erbaugh, J. (1961). An
rventory for measuring depression. Archives of General Psychiatry, 4, 561-571.

Blatt, S.J., Brenneis, C.B., Schimek, J .G., & Glick, M. (1976). Normal
evelopment and psychopathological impairment of the concept of the object on the
.orschach. Journal of Abnormal Psychology, 85, 363-373.

Block, J. (1965). The challenge of response sets: Unconfoundingpreaning,
gqpiescence, and social desirability in the MMPI. New York: Appleton-Century—Crofts.

Blumberg, E.M., West, P.M., & Ellis, F .W. (1956). MMPI ﬁndings in human
mcer. In G.S. Welsh & W.G. Dahlstrom (Eds), Basic readings on the MMPI in
sychology and medicine (pp. 452-460). Minneapolis MN: University of Minnesota
ress.

Bonanno, G.A., & Singer, J .L. (1990). Repressive personality style: Theoretical
1d methodological implication for health and pathology. In J. L. Singer (Ed),
,epression and dissociation: Implications for personality theory, psychopathology, and
aaﬂ (pp. 435-470). Chicago: The University of Chicago Press.

Breznitz, S. (1983). Denial versus hope: Concluding remarks. In S. Breznitz
3d), The denial of stress (pp. 297-302). New York: International Universities Press.

Bruehl, S., McCubbin, J .A., Wilson, J .F., Montgomery, T., Ibarra, P., & Carlson,
LR. (1994). Coping styles, opioid blockade, and cardiovascular response to stress.
upmal of Behavioral Medicine, 17, 25-40.

Butcher, J .N., Graham, J .R., Williams, C.L., & Ben-Porath, Y.S. (1990).
lavelopment and use of the MMPI-2 content scales. Minneapolis, MN: University of
linnesota Press.

Carver, C.S., & Matthews, K.A. (1989). An overview of issues in psychometric
ssessment. In N Schnedierman, SM. Weiss, & P.G. Kaufmann (Eds), Handbook of
gsearch methods in cardiovascular behavior medicine (pp.485-494). New York: Plenum
ress.

Clum, G. A., & Clum, J. (1973). Choice of defense mechanisms and their
:lationship to mood level. Psychological Reports, 32, 507-510.

Cohen, J. (1990). Things I have learned (so far). American Psychologist, 45,
304-1312.

Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155-159.

 

 

 

84

Colligan, RC. (1985). History and development of the MMPI. Psychiatric
mnals 15(9). 524-533.

Connor, ST. (1986). Measurement of denial in the terminally ill: A critical
:view. The Hospice Journal, 2, 51-68.

Conte, H.R., & Plutchik, R. (Eds). (1995). Egg defenses: theory and
lgasmement: Einstein Psychiatry Publication No. 10. New York: John Wiley & Sons.

Cooper, C., & Kline, P. (1982). A validation of the Defence Mechanism
1ventory. British Journal of Medical Psycholpgy, 55, 209-214.

Cramer, P. (1988). The Defense Mechanism Inventory: A review of research and
iscussion of the scales. Journal of Personality Assessment, 52, 142-164.

Cramer, P. (1991). The develgpment of defense mechanisms: Theory, research,
nd assessment. New York: Springer-Verlag.

Crowne, D.P., & Marlowe, D. (1964). The approval motive: Studies in evaluative
ependence. New York: John Wiley & sons

Dahlstrom, W.G., Welsh, G.S., & Dahlstrom, LE. (1972). An MMPI handbook:
[01. I. Clinical interpretation (Rev. ed.). Minneapolis, MN: University of Minnesota
’ress.

Dahlstrom, W.G., Welsh, G.S., & Dahlstrom, LE. (1975). An MMPI handbook:
(01. 11. Research applications (Rev. ed.). Minneapolis, MN: University of Minnesota
’ress.

Delistraty, D.A., Greene, W.A., Carlberg, K.A., Raver, K.K. (1992).
lardiovascular reactivity in Type A and B males to mental arithmetic and aerobic
xercise at an equivalent oxygen uptake. Psychophysiology, 29, 264-271.

Dorpat, T. L. (1985). Denial and defense in the therapeutic situation. New York:
ason Aronson.

Eagle, M. (1984). Recent developments in psychoanalysis: A critical evaluation.
Zambridge, MA: Harvard University Press.

Edelstein, E.L., Nathanson, D.L., & Stone, A.M. (Eds). (1989). Denial: A
Lariﬁcation of concepts and research. New York: Plenum Press.

Edwards, AL. (1957). The social desirability variable in personality assessment
pd research. New York: Dryden Press.

Eysenck, H.J. (1994). Neuroticism ant the illusion of mental health. American
stcholagpist, 4% 971-972.

Fairbairn, W.R.D. (1994). Psychoanalytic studies of the personality. London:
{outledge (Original work published 1952)

Fenichel, O. (1996). The psychoanalytic theory of neurosis. New York: W.W.
\lorton. (Original work published 1945).

Fine, R. (1975). Ppsychoanalytic psychology. New York: Jason Aronson.

Fordyce, W.E. (1956). Social desirability in the MMPI. Journal of Consulting
?sycholcrgy, 20(3), 171-175.

Fowler, C., Hilsenroth, M.J., & Handler, L. (1995). Early memories: An
:xploration of theoretically derived queries and their clinical utility. Bulletin of the
Awning Clinic, 59(1), 79-98.

Freud, A. (1946). The ego and the mechanisms of defence. (C. Baines, Trans).
\lew York: International Universities Press.

 

 

 

85

Freud, S. (1959). Inhibitions, symptoms and anxiety. In J. Strachey (Ed. and
rans.), The standard edition of the complete psychological wogs of Sigmund Freud
101. 20, pp. 77-175). London: Hogarth Press. (Original work published 1926 [1925]).

Fricke, B.G. (1956). Response set as a suppressor variable in the OAIS and
[IVIPL Journal of Consulting Psychology, 20(3), 161-169.

Garb, H. N. (1998). Recommendations for training in the use of the Thematic
,pperception Test (TAT). Professional Psychology: Research and Practice, 29(6), 621-
22.

Gillum, R., Leon, G.R., Kamp, J., & Becerra-Aldama, J. (1980). Prediction of
udiovascular and other disease onset and mortality from 30-year longitudinal MMPI
ata. Journal of Consulting and Clinical Psychology, 48, 405-406.

Gleser, G.C., & Ihilevich, D. (1969). An objective instrument for measuring
efense mechanisms. Journal of Consulting and Clinical Psychology, 33 (1),51-60.

Gleser, G.C., & Sacks, M. (1973). Ego defenses and reaction to stress: A
alidation study of the Defense Mechanisms Inventory. Journal of Consultingand
21inical Psychology, 40, 181-187.

Goleman, D. (1995). Emotional Intelligence. New York: Bantam.

Gordon, N.G., & Brackney, BE. (1979). Defense mechanism preference and
irnensions of psychopathology. Psychological Reports, 44, 188-190.

Graham, J .R. (1993). MMPI-2: Assessing personality aanychopathology (2nd
d.). New York: Oxford University Press.

Greene, R.L. (1999). Advanced Interpretation of the MMPI—2. Unpublished
franuscript, Paciﬁc Graduate School of Psychology.

Haan, N. (1977). Coping and defending: Processes of self-environment
»rganization. New York: Academic Press.

Hanley, C. (1957). Deriving a measure of test-taking defensiveness. Journal of
Eonsultirg Psycholggy, 21(5), 391-397.

Hathaway, S.R., & McKinley, J .C. (1989). Manual for administration and scoring
he Minnesota Multiphasic Personality Inventory-2. Minneapolis, MN: University of
Ainnesota Press.

Haythornthwaite, J .A., Pratley, R.E., & Anderson, DE. (1992). Behavioral stress
rotentiates the blood pressure effects of a high sodium intake. Psychosomatic Medicine,
34, 231-239.

Hibbard, S., Hilsenroth, M.J., Hibbard, J .K., & Nash, MR. (1995). A validity
:tudy of two projective object representations measures. Psychological Assessment, 7,
132-439.

Heilbrun, A.B., Jr. (1961). The psychological signiﬁcance of the MMPI K scale
n a normal population. Journal of Consulting Psychology. 25, 486-491.

Hohnes, D.S. (1990). The evidence for repression: An examination of sixty years
)f research. In J. L. Singer (Ed.), Repression and dissociation: Implications for
)ersonality theory, psychopathology and healph (pp. 85-102). Chicago: The University
)f Chicago Press.

Houston, B.K. (1973). Viability of coping strategies, denial, and response to
;tress. Journal of Persarlality, 41, 50-58.

86

 

Ihilevich, D., & Gleser, GO (1986). Defense mechanisms: Their classiﬁcation,
prrelates, and measurement with the Defense Mechanisms Inventory. Owosso, MI: DMI
ssociates.

Ihilevich, D., & Gleser, GO (1995). The Defense Mechanisms Inventory: Its
:velopment and clinical applications. In HR. Conte & R. Plutchik (Eds), Egg
:fenses: Theory and measurement: Einstein Psychiatry Publication No. 10 (pp. 221-

17). New York: John Wiley & Sons

Joiner, T.E., Jr. (1994). On denial and defensiveness in experimental psychology.
Lrnerican Psychologist,(49)11, 973-974.

Jorgensen, R.S., Schreer, G.E., Baskin, L., & Kolodziej, M. (1992). Denial and
re discrepancy between heart rate and reported negative affect: A study of convergent
1d discriminant validity. Psychotherapy and Psychosomatics, 58, 202-207.

Karon, BR (1968). The Thematic Apperception Test (TAT). In A.I. Rabin (Ed),
LOj ective techniques in personality assessment (pp. 85-120). New York: Springer
ublishing Company.

Katkin, E.S., Dermit, S. & Wine, S.K.F. (1993). Psychophysiological assessment
f stress. In L. Goldberg & S. Breznitz (Eds), Handbook of stress: Theoretical and
linical aspects (2nd Ed.), (pp. 142-157). New York: Free Press.

Kavanagh, G.G. (1985). Changes in patients' object representations during
sychoanalysis and psychoanalytic psychotherapy. Bulletin of the Menningar Clinic, 49,
46-564.

Kemberg, O. (1976). Object-relations theory and clinical psychoanalysis. New
York: Jason Aronson.

Kissen, M. (Ed.). (1986). Assessing object relations phenomena. Madison, WI:
ntemational University Press.

Lair, C.V., & King GD. (1976). MMPI proﬁle predictors for successful and
:xpired open heart surgery patients. Journal of Clinical Psychology, 32, 51-54.

Lanyon, R.I., & Lutz, R.W. (1984). MMPI discrimination of defensive and
tondefensive felony sex offenders. Journal of Consulting and Clinical Psychology, 52(5),
:41-843.

Larson, D.G., & Chastain, R.L. (1990). Self-concealment: Conceptualization,
measurement, and health implications. Journal of Social and Clinical Psychology, 9, 43 9-
145.

Lazarus, R.S., & Alfert, E. (1964). Short-circuiting of threat by experimentally
rltering cognitive appraisal. Journal of Abnormal and Social Psychology, 69, 195-205.

Lazarus, R.S., & Folkman, S. (1991). The concept of coping. In A. Monat & R.S.
gazarus (Eds), Stress and coping: An anthology (3rd ed.), (pp. 189-206). New York:
Columbia University Press.

Levitt, BE. (1980). The psychology of anxiety. (2nd ed.). Hillsdale, NJ:

Lawrence Erlbaum Associates.

Levitt, EB. (1989). The clinical application of MMPI special scales. Hillsdale,
\IJ: Lawrence Erlbaum Associates.

Little, K.B., & Fisher, J. (1958). Two new experimental scales of the MMPI.
Journal of Consulting Psycholgy, 22, 305-306.

87

Lubin, B., Larsen, R.M., & Matarazzo, J .D. (1984). Patterns of psychological test
ge in the United States: 1935-1982. American Psychologist, 39, 451-454.

Luborsky, L. (1984). Principles of psychoanalytic psychotherapy: A manual for
portive-expressive treatment. Basic Books

Mandler, G.M., Mandler, J .M., Kremen, I., & Sholiton, RD. (1961) The response
hreat: Relations among verbal and physiological indices. Psychological Monographs,
(9, Whole No. 513).

Mayman, M. (1967). Object-representations and object-relationships in
rschach responses. Journal of Projective Techniques, 31, 17-24.

McGrath, R.E., & O'Malley, W.B. (1986). The assessment of denial and physical
nplaints: The validity of the Hy scale and associated MMPI signs. Journal of Clinical
fchology, 42, 754-760.

McKinley, J .C., Hathaway, S.R., & Meehl, RE. (1956). The K scale. In G.S.

:lsh & W.G. Dahlstrom (Eds), Basic readings on the MlVfPI in psychology and
dicine (pp. 112—123). Minneapolis MN: University of Minnesota Press.

Mead, PG, (197 0). The effect of orientation passages on patient stress prior to
rtistry. The Psychological Record, 20, 479-488.

Meehl, PE. (1956). The dynamics of "structured" personality tests. In G.S.

:lsh & W.G. Dahlstrom (Eds), Basic readings on the MMPI in psychology and
:dicine (pp. 5-11). Minneapolis MN: University of Minnesota Press. (Original work
blished 1945) ‘

Meehl, P.E., & Hathaway, SR. (1956). The K factor as a suppressor variable in
: MMPI. In G.S. Welsh & W.G. Dahlstrom (Eds), Basic readings on the MMPI in
ychology and medicine (pp. 12-40). Minneapolis MN: University of Minnesota Press.
riginal work published 1946)

Myers, J .K., & Bean, LL. (1968). A decade later: A follow-up of social class and
:ntal illness. New York: John Wiley and Sons.

Nigg, J .T., Lohr, N.E., Westen, D., Gold, L.J., & Silk, KR. (1992). Malevolent
ject representations in borderline personality disorder and major depression. Journal of
)normal Psychology, 101, 61-67.

Pancheri, P., Bellaterra, M., Matteoli, D., Cristorari, M., Polizzi, C., & Puletti, M.
.978, March). Infarct as a stress agent: Life history and personality characteristics in
[proved versus not-improved patients after severe heart attack. Journal of Human
peg, 16-22, 41-42.

Peglar, M., & Borgen, F .H. (1984). The defense mechanisms of coronary
Ltients. Journal of Clinical chhology, 40, 669-679.

Pennebaker, J .W. (1985). Traumatic experience and psychosomatic disease:
Kploring the roles of behavioral inhibition, obsession and conﬁding. Canadian
gychologist. 26, 82-95.

Pennebaker, J .W., & Beal, SK. (1986). Confronting a traumatic event: Toward
1 understanding of inhibition and disease. Journal of Abnormal Psycholagy, 95, 274-
31.

 

 

 

 

Pennebaker, J .E., Kiecolt-Glaser, J .K., & Glaser, R. (1988). Disclosure of
aumas and immune function: Health implications for psychotherapy. Journal of

gsulting and Clinical Psychology, 56, 239-245.

88

Pennebaker, J .W., & O'Heeron, R. (1984). Conﬁding in others and illness rate
Long spouses of suicide and accidental-death victims. Journal of Abnormal
ychology, 93, 473-476.

Persky, V.W., Kempthome-Rawson, J ., & Shekelle, RB. (1987). Personality and
k of cancer: 20-year follow-up of the Western electric study. Psychosomatic Medicine,
, 435-449.

Piotrowski, C., & Keller, J .W., (1989). Psychological testing in outpatient mental
11th facilities: A national survey. Professional Psychological: Research and Practice,

, 423-425.

Plutchik, R. (1995). A theory of ego defenses. In HR. Conte & R. Plutchik
:ls.), Ea) defenses: Theory and measurement: Einstein Psychiatry Publication No. 10
). 13-38). New York: John Wiley & Sons.

Robins, L.N., Locke, B.Z., & Regier, DA. (1991). An overview of psychiatric
:orders in America. In L.N. Robins & D.A. Regier (Eds), Psychiatric disorders in
nerica: The epidemiologic catchment area study (pp. 328-3 66). NY: Free Press.

Schonpﬂug, W. (1986). Behavior economics as an approach to stress theory. In
.H. Appley & R. Trumbull (Eds), Dynamics of stress: Physiological, psychological,

d social perspectives (pp. 81-98). New York: Plenum Press.

Schuldberg, D. (1992). Ego-strength revised: A comparison of the MMPI-2 and
MPI-l versions of the Barron Ego-Strength Scale. Journal of Clinical Psychology, 48.
0-505.

Schwartz, GE. (1990). Psychobiology of repression and health: A systems
proach. In J. L. Singer (Ed), Repression and dissociation: Implications for personality
:ory, psychopathology, and health (pp. 405-434). Chicago: The University of Chicago
ess.

Selye, H. (1991). History and present status of the stress concept. In A. Monat &
S. Lazarus (Eds), Stress and coping: An anthology (3rd ed), (pp. 21-35). New York:
)lumbia University Press.

Shanan, J. (1989). The place of denial in adult development. In E.L. Edelstein,
L. Nathansosn, & A.M. Stone (Eds), Denial: A clariﬁcation of concepts and research
p. 107-188). New York: Plenum Press.

Shapiro, D.L. (1991). Forensic psychological assessment: An integrative
proach. Boston: Allyn and Bacon.

Shedler, J ., Mayman, M., & Manis, M. (1993). The illusion of mental health.
mrican Psychologist,(48)11, 1117-1131.

Shedler, J ., Mayman, M., & Manis, M. (1994). More illusions. American
yhologist,(49)l 1, 974-976.

Singer, J .L., & Sincoff, J .B. (1990). Summary: Beyond repression and the
:fenses In J. L. Singer (Ed), Repression and dissociation: Implications for personality
gory, psychopathology, and health (pp. 471-496). Chicago: The University of Chicago
'ess.

Sjoback, H. (1973). The psychoanalytic theory of defensive processes. New
ork: John Wiley & Sons.

 

 

 

 

 

89

 

Sjoback, H. (1991). Defence, defence and defence: How do we measure
:fence? In M. Olff, G. Godaert, & H. Ursin (Eds), Quantiﬁcation of human defence
,echanisms (pp. 4-21). Berlin, Germany: Springer-Verlag.

Smith, BE. (1959). Defensiveness, insight and the K scale. Journal of
onsulting Psychology, 23(3), 275-277.

Smith, TE. (1993). Measurement of object relations: A review. Journal of
sychotherapy Practice and Research, 2(1), 19-37.

Stricker, G., & Healey, B.J. (1990). Projective assessment of object relations: A
view of the empirical literature. Psychological Assessment: A Journal of Consulting
1d Clinical Psychology,_; 219-230.

Tappan, D.V., & Weybrew, BB. (1982, April). Relationship of personality
ctors and some social habits to cardiovascular risk in submariners. Aviation, Space,
rd Environmental Medicine, 383-389.

Taylor, J .B., Carithers, M., & Coyne, L. (1976). MMPI performance, response
:t, and the "self-concept hypothesis". Journal of consulting and Clinical Psychology, 44,
51-3 62.

Taylor, J .G. (1987). Psychosomatic medicine and contemporary psychoanalysis.
ladison, CT: International University Press.

Taylor, S.E., & Brown, J .D. (1988). Illusion and well being: A social
:ychological perspective on mental health. Psychological Bulletin, 103, 193-210.

Taylor, S.E., & Brown, J .D. (1994). "Illusion" of mental health does not explain
)sitive illusions. American PsychologsL(49)l 1, 972-973.

Temoshok, L. (1993). Emotions and health outcomes: Some theoretical and
ethodological considerations. In H.C. Traue & J .W. Pennebaker (Eds), Emotion
hibition apd health (pp. 247-256). Seattle WA: Hogrefe & Huber Publishers.

Thompson, AB. (1986). An object relational theory of affect maturity:
pplications to the Thematic Apperception Test. In M. Kissen (Ed), Assessing object
lations phenomena (pp. 207-224). Madison, WI: International Universities Press.

Trimboli, F., & Kilgore, RB. (1983). A psychodynamic approach to MMPI
terpretation. Journal of Personality assessment, 47, 614-626.

Staff. (1999, September 27). Undergrads used unfairly in grad research. The State
mp. 4A

Vaillant, GE. (1990). Repression in college men followed for half a century. In
L. Singer (Ed), Repression and dissociation: Implications for personality theory,
ychopathology, and health (pp. 259-274). Chicago: The University of Chicago Press.

Van Egeren, LR (1989). The analysis of continuous data. In N. Schneiderman,
M. Weiss, & P.G. Kaufrnann (Eds), Handbook of research methods in cardiovascular
:havioral medicine (pp. 589-614). New York: Plenum Publishing.

Watson, G.C., Plemel, D., Vassar, P., Manifold, V., Kucala, T., & Anderson, D.
987). The comparative validities of six MMPI repression scales. Journal of Clinical
sychology, 43, 472-477.

Wegner, D.M., & Pennebaker, J .W. (1993). Changing our minds: An introduction

mental control. In D.M. Wegner & J .W. Pennebaker (Eds), Handbook of mental
map] (pp.1-12). Englewood Cliffs, NJ: Prentice Hall.

 

 

9O

 

Weinberger, DA. (1990). The construct validity of the repressive coping style.
1 J .L. Singer (Ed), Repression and dissociation : Implications for personality theory,
sychopathologv, and health (pp. 337-3 86). Chicago: University of Chicago Press.

Weinberger, D., Schwartz, G., & Davidson, R. (1979). Low anxious, high-
1xious, and repressive coping styles: Psychometric patterns and behavioral and
iysiological responses to stress. Journal of Abnormal Psychology, 88, .

Weinstein, J. Averill, J .R., Opton, E.M., & Lazarus, RS. (1968). Defensive style
1d discrepancy between self-report and physiological indexes of stress. Journal of
ersonality and Social Pchhology, 10, 406-413.

Welsh, GS. (1956). Factor Dimensions A and R. In G.S. Welsh & W.G.
ahlstrom (Eds), Basic readings on the MMPI in psychology and medicine (pp. 264-
39). Minneapolis MN: University of Minnesota Press.

Westen, D. (Ed). (1990). Social cognition and object relations scale (SCORS):
lanual. for coding TAT data. (Available from D. Westen, Cambridge Hospital, Harvard
[edical School).

Westen, D. (1991a). Social cognition and object relations. Psychological
ulletin, 109, 429-455.

Westen, D. (1991b). Clinical assessment of object relations using the TAT.
)urnal of Personality Assessment, 56, 56-74.

Westen, D., Lorh, N., Silk, K., Gold, L., & Kerber, K. (1990). Object relations
1d social cognition in borderlines, major depressives, and normals: A thematic
)perception test analysis. Psychological Assessment: A Journal of Consulting and
linical Psychology, A 355-364.

Wilson, J .F . (1982). Recovery from surgery and scores on the Defense
[echanisms Inventory. Journal of Personality Assessment, 46, 312-319.

 

 

 

 

 

9’1

 

able 1

orrelations between SCORS Assessment and Physiological Indices of Psychological

 

 

 

 

 

ealth/Distpess
PHYSIOLOGICAL INDEX
Rate Pressure Product
Average Stress Change Most Extreme Change
CORS
fr Scale .11 (p=.20) .oo (p=.50)
.T+EI Scale .10 (p=.23) -.03 (p=.40)
1 Scale .05 (p=.35) -.05 (P=.36)
T PHYSIOLOGICAL INDEX
Mean Arterial Pressure
Average Under Stress Change From Baseline
CORS
fr Scale .14 (p=.15) .06 (p=.33)
.T+EI Scale .20 (p=.07) .07 (p=.31)
.1 Scale .17 (p=.11) .05 (p=.37)

 

92

 

 

able 2

orrelations between MMPI-2 Subscales and Physiological Indices of Psychological

 

 

 

 

 

palth/Distress
T PHYSIOLOGICAL INDEX
Rate Pressure Product
lMPI-2 Average Stress Change Most Extreme Change
7elsh Anxiety .09 (p=.26) .17 (p=.l 1)
,go Strength -.05(p=.37) -.07 (p=.30)
iollege Maladjustment .11 (p=.22) .20 (p=.07)
1 Scale -.02 (p=.44) -.10 (p=.24)
)enial Scale .09 (p=.25) .04 (p=.39)
PHYSIOLOGICAL INDEX
Mean Arterial Pressure
Average Under Stress Change From Baseline
Velsh Anxiety .12 (p=.20) .20 (p=.08)
Ego Strength -.12 (p=.18) -.16 (p=.12)
Zollege Maladjustment .17 (p=.11) .11 (p=.21)
C Scale -.17 (p=.11) -.20 (p=.07)
)enial Scale -.04 (p=.39) -.03 (p=.42)

 

Sote. None of the MMPI-2 scales were signiﬁcantly correlated to the baseline measure,
herefore it was not necessary to partial out baseline effects.

93

 

 

 

Table 3
30rrelations between MMPI-2 Scales of Ppychological Health and Physiological Indices

1f Psychological Health/Distress While Controllingior Denial

 

PHYSIOLOGICAL INDEX
Rate Pressure Product
Average Stress Change Most Extreme Change

 

 

vIMPI-2

Nelsh Anxiety .26 (p=.03) .33 (p=.00)

Ego Strength -.12 (p=.19) -.12 (p=.20)

Zollege Maladjustment .23 (p=.05) .31 (p=.01)

PHYSIOLOGICAL INDEX
Mean Arterial Pressure

VlMPI-2 Average Under Stress Change From Baseline
Welsh Anxiety .14 (p=.16) .28 (p=.01)

Ego Strength -.13 (p=.18) -.18 (p=.10)

College Maladjustment .19 (p=.08) .13 (p=.19)

 

94

 

,‘able 4

Iorrelations between MMPI-2 Subscales and SCORS Scales

 

 

 

 

SCORS SCALE
AMPI-2 Scale Affect Tone AT and EI Emotional Investment
Nelsh Anxiety -.18 (p=.09) -.12 (p=.19) -.O3 (p=.42)
Ego Strength .12 (p=.19) -.06 (p=.33) -.18 (p=.10)
vIaladjustment -.18 (p=.09) -.11 (p=.21) -.02 (p=.45)
SCORS SCALE
.VIMPI-Z Scale Affect Tone AT and EI Emotional Investment
i< Scale .28 (p=.02) .16 (p=.12) .00 (p=.49)
Denial Scale .33 (p=.01) .30 (p=.01) .17 (p=.12)

 

Note. The Maladjustment scale refers to the College Maladjustment scale.

95

Table 5
Correlationp between MMPI-2 Scales of Global Psychological Health and SCORS Scales

while Controlﬂg for Denial

 

SCORS SCALE
Affect Tone AT and EI Emotional Investment
Welsh Anxiety .13 (p=.17) .20 (p=.07) .17 (p=.11)
Ego Strength -.09 (p=.25) -.3l (p=.01) -.35 (p=.00)
Maladjustment .05 (p=.35) .13 (p=.18) .13 (p=.18)

 

Note. The Maladjustment scale refers to the College Maladjustment scale. Also, recall
that on the Ego Strength scale, low scores indicate more distress.

96

 

Table 6
Mean Differences on Physiological Reactivity Scores in Healthy versus Distressed

Groups (based on a Mean Split of Affect Tones scores)

 

Rate Pressure Product

 

Average Stress Change Most Extreme Change
M SD M SD
Distressed (n=30) 13.07 9.54 29.99 14.18
Healthy (n=25) 14.68 13.90 29.55 20.03

 

Mean Arterial Pressure

 

Average Under Stress Change From Baseline
M SD M SD
Distressed (n=30) * * 10.49 6.16
Healthy (n=25) * * 9.17 6.47

 

Ma, The RPP baseline physiological scores for the two groups were nearly identical and
there was a nonsigniﬁcant difference between them, t(53)=-.3 6, p=.7l7, two-tailed.
Distressed versus Healthy groups were based on a mean split of Affect Tone scores as
follows: Distressed = AT<2.79; Healthy = At>2.80.

97

Table 7

Mean Differences on Physiological Reactivig Scores in Healthy, Distressed, and

Undetermined Groups (based on Affect Tones scores)

 

Rate Pressure Product

 

Average Stress Change Most Extreme Change
M SD M SD
Distressed (n=24) 14.14 9.81 31.72 15 .02
Healthy (n=l4) 20.00 14.80 35.28 20.18
Undetermined (n=17) 8.22 8.57 22.56 14.96

 

Mean Arterial Pressure

 

Average Under Stress Change From Baseline
M SD M SD
Distressed (n=24) 94.27 8.26 10.47 6.36
Healthy (n=l4) 98.52 9.04 12.34 4.59
Undetermined (n=17) 93.16 7.25 7.05 6.56

 

1% There were nonsigniﬁcant differences between any of the group baseline means.
The Undetermined group's RPP Average Stress Change and Most Extreme Change means
were signiﬁcantly different than those mean scores for the Healthy Group (p<.05). The
Undetermined group's RPP Average Stress Change mean was signiﬁcantly different than
that mean score for the Distressed Group (p<.05). The difference between the Extreme
Scores approached signiﬁcance (p=.06).

The undetermined group's MAP Change From Baseline mean score was signiﬁcantly
different from the Healthy group's mean score (p<.05). None of the means were
signiﬁcantly different between the healthy versus distressed group.

Distressed, Healthy, and Undetermined groups were based on Affect Tone scores as
follows: AT<=2.63 Distressed; AT>=2.94=Healthy; At between 2.64 and 2.93 were

Undetermined.

98

 

 

 

Table 8
Mean Differences on Physiological Reactivity Scores in Healthy, Distressed, and Illusory

Groups (Based on Affect Tone scores and Welsh Anxiety Scores)

 

Rate Pressure Product

 

Average Stress Change Most Extreme Change
M SD M SD
Distressed (n=12) 10.10 7.74 26.04 14.57
Healthy (n=16) 14.16 10.72 30.25 13.65
Illusory (n=22) 12.92 10.69 28.45 15.45

 

Mean Arterial Pressure

 

Average Under Stress Change From Baseline
M SD M SD
Distressed (n=12) 94.83 7.14 11.33 6.32
Healthy (n=16) 97.15 9.23 9.51 6.15
Illusory (n=22) 92.67 8.12 9.01 6.1 3

 

Note. There were nonsigniﬁcant differences between any of the group baseline means.
Further, there were no signiﬁcant differences between any of the group means.

Distressed, Healthy, and Illusory groups were based on Affect Tone scores and Welsh

Anxiety scale scores as follows: Distressed: A>=5 8, AT<=2.85; Healthy: A<58 AT>2.85;
Illusory: A<58, AT<=2.85

99

Table 9
Mean Differences on Physiological Reactivity Scores in Healthy, Distressed, and Illusory

Groups (Based on Affect Tone scores and Welsh Anxigy Scores)

 

Rate Pressure Product

 

Average Stress Change Most Extreme Change
M SD M SD
Distressed (n=10) 11.69 7.50 29.44 13.45
Healthy (n=11) 15.05 11.00 27.88 10.29
Illusory (n=20) 13.77 10.52 30.27 14.86

 

Mean Arterial Pressure

 

Average Under Stress Change From Baseline
M SD M SD
Distressed (n=10) 93.48 6.69 12.83 5.65
Healthy (n=11) 97.48 9.95 11.35 3.66
Illusory (n=20) 93.47 8.34 9.32 6.20

 

1% There were nonsigniﬁcant differences between any of the group baseline means
(however, the difference between the MAP baseline measures for the healthy versus
distressed group approached signiﬁcance, t(19)=1.85, p=.08). Further, there were no
signiﬁcant differences between any of the group means

Distressed, Healthy, and Illusory groups were based on Affect Tone scores and Welsh
Anxiety scale scores as follows: Distressed: A>=59, AT<=2.75; Healthy: A<59 AT>=3.0;
Illusory: A<59, AT<=2.75. These results differ from Table 8 by using different cut-off
scores on the AT scale. That is, 11 participant's with AT scores around the mean (i.e.,
AT>2.75 and AT<3.0) were not used in the analysis.

100

 

 

 

Table 10

Mean Differences on Physiological Reactivity Scores in Healthy, Distressed, and Illusory

Groupp (Based on the Combination of Affect Tone and Emotional Intelligence scores, as

well as Welsh Anxiety Scores)

 

 

Rate Pressure Product

 

 

 

Average Stress Change Most Extreme Change
M SD M SD
Distressed (n=l4) 13.49 12.16 22.25 21.82
Healthy (n=15) 12.74 10.87 27.60 12.45
Illusory (n=21) 13 .63 10.97 29.07 15 .96
Mean Arterial Pressure
Average Under Stress Change From Baseline
M SD M SD
Distressed (n=l4) 95.41 7.11 11.16 6.31
Healthy (n=15) 97.11 9.30 10.61 5.31
Illusory (n=21) 93.54 8.70 8.87 6.47

 

Note. Though not shown above, the Illusory group had a signiﬁcantly lower RPP
baseline than the healthy group. There were no other signiﬁcant differences between any

of the group means.

Distressed, Healthy, and Illusory groups were based on the combination of Affect Tone
and Emotional Intelligence scores (EI) scores, as well as Welsh Anxiety scale scores as
follows: Distressed: A>=59, AT&EI<=2.6; Healthy: A<59 AT&EI>2.6; Illusory: A<59,

AT&EI<=2.6.

10’1

 

 

Table 11

Correlational Analyses of SCORS AT Scores with Various Study Scales

 

SCORS AT Scale
Denial Scales
MMPI-2 Denial Scale .33 (p=.01)
MMPI-2 L Scale .24 (p=.04)
MMPI-2 K Scale .28 (p=.02)
DMI REV Scale .26 (p=.03)
Marlowe Crowne Scale .27 (p=.02)

 

’102

 

Appendix A
Defense Mechanism Inventory Example

DMI Story:
You are waiting for the bus at the edge of the road. The streets are wet and

muddy after the previous night’s rain. A car sweeps through a puddle in front of you,
splashing your clothing with mud.

What would your actual response he?

I would note the car’s license number so that I could track down that careless driver.
I’d wipe myself off with a smile.

I’d yell curses after the driver!

I would scold myself for not having at least worn a raincoat.

I’d shrug it off; after all things like that are unavoidable.

SAPP’NE“

103

 

Appendix B
Little and Fisher Denial scale
Scored if False:
MMPI-1 #6; MMPI-2 #7: I like to read newspaper articles on crime.
MMPI-1 # 12; MMPI-2 # 14: I enjoy detective or mystery stories.

MMPI-1 and 2 # 26: feel that it is certainly best to keep my mouth shut when I'm in
trouble.

MMPI-1 #30; MMPI-2 #29: At times I feel like swearing.

MMPI-1 #71; MMPI-2 #58: I think a great many people exaggerate their misfortunes in
order to gain the sympathy and help of others.

MMPI-1 #89; MMPI-2 #76: It takes a lot of argument to convince most people of the
truth.

MMPI-1 #93; MMPI-2 #81: I think most people would lie to get ahead.

MMPI-1 #109; MMPI-2 #98: Some people are so bossy that I feel like doing the opposite
of what they request, even though I know they are right.

MMPI-1 #124; MMPI-2 #110: Most people will use somewhat unfair means to gain
proﬁt or advantage rather than to lose it.

MMPI-1 #129: Often I can't understand why I have been so cross and grouchy.
MMPI-2 #116: Often I can't understand why I have been so irritable and grouchy.

MMPI-1 #136: I commonly wonder what hidden reason another person may have for
doing something nice for me.

MMPI-2 #124: I often wonder what hidden reason another person may have for doing
something nice for me.

MMPI-1 #141: My conduct is largely controlled by the customs of those about me.
MMPI-2 #129: My conduct is largely controlled by the behavior of those around me.

MMPI-1 #147; MMPI-2 #135: I have often lost out on things because I couldn't make up
my mind soon enough.

MMPI-1 #162: I resent having anyone take me in so cleverly that I have had to admit that
it was one on me.

104

 

Appendix B (cont’d)

MMPI-2 #151: I resent having anyone trick me so cleverly that I have to admit I was
fooled.

MMPI-1 #170; MMPI-2 #157: What others think of me does not bother me.
MMPI-1 #172; MMPI-2 #161: I frequently have to ﬁght against showing that I am
bashful.

MMPI-l #180; MMPI-2 #167: I ﬁnd it hard to make talk when I meet new people.
MMPI-1 #201; MMPI-2 #185: I wish I were not so shy.

MMPI-1 #213; MMPI-2 #193: In walking I am very careful to step over Sidewalk cracks.
MMPI-1 #234; MMPI-2 #213: I get mad easily and then get over it soon.
MMPI-1 #265; MMPI-2 #241: It is safer to trust nobody.

MMPI-1 #267; MMPI-2 #243: When in a group of people I have trouble thinking of the
right things to talk about.

MMPI-1 #279; MMPI-2 #253: I drink an unusually large amount of water every day.

MMPI-l #289; MMPI-2 #262: I am always disgusted with the law when a criminal is
freed through the arguments of a smart lawyer.

MMPI-1 #292; MMPI-2 #265: I am likely not to speak to people until they speak to me.

Scored if true
MMPI-1 #253; MMPI-2 #230

I can be friendly with people who do things which I consider wrong.

105

Appendix C

Phrase Association Task

Neutral Theme Block (Set 1):
The horses worked well together
The library purchased more books
The artist ﬁnished the painting
The new movie opens this Friday
The plumber repaired the sink
The woman walked her dog

The airline tickets were conﬁrmed
The maihnan arrived on time

The ﬂowers blossomed

The waitress was friendly

Aggressive Theme Block:

The father was convicted for torturing his son
The student was attacked by a gang

He suddenly struck his teacher

She shot her boyﬁiend

The mother scalded her infant in the bath
He spit in his mother's face

The driver ran her off the road

The jury voted on the death penalty

The boy pulled the wings off of the butterﬂy
The girl threw mud on her friends dress

Dependency Theme Block:

The mother bear deserted the baby cubs

He needed help with his homework

The father neglected the sick child

She pleaded with her mother

The bank refused to lend the couple money
The mother sent her daughter to bed hungry
His brother refused to help

The family had to beg for food

The puppy was unable to suckle his mother
The woman was so exhausted She needed help walking home

106

 

Appendix C (cont’d)

Sexual Theme Block:

Prostitutes do anything men desire

He enjoys sleeping with men

She propositioned the waiter

After the operation he was impotent
The two women embraced and kissed in public
His roommate made a pass at him

She purchased the condoms yesterday
The couple conceived their ﬁrst child
The results came back from the HIV test
He took off her clothes

Neutral Theme Block (Set 2):

The new car gets good gas mileage

The builder completed the house

The museum opens at noon

The weather was pleasant

The team ﬁnished practice early

The radio gave free passes to the concert
The new shoes ﬁt comfortably

The picture hung on the wall

The new carpeting accented the furniture
The air smelled sweet

107

 

 

Appendix D
The Calculation of Socio-Economic Status

The participant's socio-economic status (SES) was calculated using Hollingshead's
Two Factor Index cited in Myers and Bean (1968). Hollingshead originally used a three
factor index to calculate SES which included the occupation, education, and residential
setting of the head of household. The residential setting was dropped from the index
because it was determined that a detailed knowledge about the residential setting within the
city or suburb was necessary for this factor to be accurate. Subsequently, the education
(measured in years of completed schooling) and occupation (divided into 7 categories to
represent earning power) of the head of household became the common factors used to
assign an SES.

In this study, if the participant was married and lived independent ﬁom the parents,
either the participant or the participant's spouse, whichever would result in the highest SES,
was used in the calculation of SES. Ifthe participant lived with his or her parents or was
dependent upon them for support (as in the case of most college students), the parent that
placed the participant in the highest SES category was used as the head of the household for
calculations.

Both the education and occupation indexes are multiplied by weighting factors then
the two scores are summed. The range of scores run along a continuum from 11 to 77, with
the lowest score equating to the highest SES. The range of scores are then subdivided into
ﬁve social class categories as follows:

Range of Computed Scores Social Class
11 - 17 I
18 - 27 H
28 - 43 [II
44 - 60 IV
61 - 77 V

In this study the participant score (ﬁom 11 to 77) was used in statistical calculations to
eliminate the use of a categorical variable, and subsequently improve statistical inferences.

’108

 

Appendix E

Global health Assessment Based on TAT Responses: Explanatog Examples

Both of the stories below came from actual study participants and are in response to TAT
card 6BM.

Healthy Example

Um, this son came and told his mother that she should move out of her house and perhaps
into a nursing home--she kinda looks scared and confused, and he looks like this isn't
something that he wants to have her do, but thinks that she should. I think that she'll go,
because her son thinks that she should, and uh she'll move out. (outcome?) I think he
feels guilty about having her move out, but he thinks it's the best thing for her and so he'll
help her move, perhaps she'll move into his house or even the nursing home or something
like that. She looks like she'll go along with it, not really willingly, but her son knows
what's best for her at this point.

Unhealthy Example

Um, it looks like the old lady is his mother-~he just took her to a nursing home because he
can't care for her anymore, and she's not too happy about it. He feels bad, but he leaves
her there (what is the woman thinking?) She doesn't want to be there, she wants to be on
her own still. (outcome) Um, she ends up dying in the nursing home, and she was lonely--
she died of old age.

Discussion

In both stories the son was faced with putting his mother in a nursing home. The
ﬁrst story was judged healthy because the son's conﬂict was made apparent, but it was
clear that he was making the choice in the best interest of the mother. Further, his
mother, though apprehensive, appears to have basic trust in her son and more readily goes
along with the decision.

In the second story, the same decision is made by a son to place his mother ("the
old lady") in a nursing home. However, instead of developing an internal conﬂict about
the decision, the son "feels bad, but leaves her there." There is an element of coldness, or
lack of caring that seems to come through in the story. We further see that the mother
was lonely and died, followed with what seemed like an afterthought that she died "of old
age". This last statement seems to have been made to distance the son from possible
guilty feelings and deny responsibility for the mother's death.

Additionally in the healthy story, it seemed clear that the decision was made in the
mother's best interest. In the second story, the decision was made for the son's interest
because "he can't care for her anymore". The ﬁrst story conveyed more warmth,
compassion, caring, and trust between the characters, whereas in the second story the
characters seemed to lack affective and emotional development.

109

 

 

 

 

1.. 1 1.1,: .
.1511.

. ,
".4... 5...... 1.1..
. I 3:151
a t: (.12
.7. 1.3.11...
:64 .3
.5... x. i.

5.. I
a