Journal of Abnormal Psychology Copyright 1991 by the American Psychological Association, Inc.

1991, Vol. 100, No. 1, 98-101 0021-843X/91/$3.00

Dimensions of Perfectionism in Unipolar Depression

Paul L. Hewitt Gordon L. Flett
Brockville Psychiatric Hospital, Brockville, York University, Toronto,
Ontario, Canada and Department o f Psychiatry, Ontario, Canada
University o f Ottawa, Ottawa, Ontario, Canada

We tested the hypothesis that self-oriented perfectionism, other-oriented perfectionism, and so-
cially prescribed perfectionism are related differentially to unipolar depression. The Multidimen-
sional Perfectionism Scale was administered along with measures of depression and anxiety to 22
depressed patients, 22 matched normal control subjects, and 13 anxiety patients. It was found that
the depressed patients had higher levels of self-oriented perfectionism than did either the psychiat-
ric or normal control subjects. In addition, depressed patients and anxious patients reported higher
levels of socially prescribed perfectionism than did the normal control subjects. The results suggest
that various dimensions of perfectionism may play an important role in clinical depression.

Recently there has been renewed interest in personality fac-
tors related to depression (see Carson, 1989). Perfectionism is a
construct that has been described as a potential vulnerability
factor in depression by both psychoanalytic (Bibring, 1953) and
cognitive theorists (Beck, 1967; Kanfer & Hagerman, 1981 ). For
example, Kanfer and Hagerman contended that excessive self-
standards serve to increase the frequency and magnitude o f
failure experiences. Perfectionistic self-standards and atten-
dant failure experiences combine with self-blame or distress to
produce depression.
Investigations have confirmed the presence o f an association
between perfectionism and subclinical depression (Hewitt &
Dyck, 1986; Hewitt & Flett, 1990); however, these studies are
limited in several respects. For example, there have been no
direct examinations of perfectionism in clinically depressed pa-
tients, so it is not known whether past findings generalize to the
experience of clinical depression. The need for direct research
is further indicated by the equivocal findings o f previous stud-
ies that have examined clinical depression and certain compo-
nents of the perfectionism construct, including high self-stan-
dards and discrepancies between the actual and ideal self
(Carver, LaVoie, Kuhl, & Ganellen, 1988; Ganellen, 1988; Strau-
man, 1989).
Finally, most previous research has been limited by a ten-
An earlier version ofthis article was presented at the annual meeting
of the Canadian Psychological Association, Ottawa, Ontario, Canada,
May 1990. This research was supported by Grant 410-89-0335 from
the Social Sciences and Humanities Research Council of Canada as
well as by a grant from the Research and Program Evaluation Commit-
tee, Brockville Psychiatric Hospital, Brockville, Ontario, Canada.
We thank Jane Baldock, Jeff Jackson, Amde Teferi, Aygodan Ugur,
and Zul WaUani for referring patients and Marjorie Cousins, Charles
Massey, Rosemary Smith, Wendy Turnbull-Donovan, and the Friend-
ship Centre, Brockville, Ontario, Canada, for their assistance. We also
express our gratitude to Norman Endler and the three reviewers for
their comments.
Correspondence concerning this article should be addressed to Paul
L. Hewitt, Department of Psychology, Elmgrove Unit, Brockville Psy-
chiatric Hospital, Brockville, Ontario K6V 5W7, Canada.
dency to view the perfectionism construct from a unidimen-
sional cognitive perspective (Burns, 1983). Recent studies have
shown that the perfectionism construct is multidimensional
and has both personal and social aspects (Frost, Marten, La-
hart, & Rosenblate, 1990; Hewitt & Flett, in press). For in-
stance, factor analyses of our Multidimensional Perfectionism
Scale (MPS; Hewitt & Flett, 1989, in press) have identified three
components of perfectionism: Self-Oriented, Other-Oriented,
and Socially Prescribed Perfectionism. According to Hewitt
and Flett (in press), self-oriented perfectionism is an intraper-
sonai dimension characterized by a strong motivation to be
perfect, setting and striving for unrealistic self-standards, fo-
cusing on flaws, and generalization o f self-standards. Self-or-
iented perfectionism may also involve a well-articulated ideal
self-schema (see Hewitt & Genest, 1990). Other-oriented per-
fectionism involves similar behaviors, but these behaviors are
directed toward others instead o f toward the self. Finally, so-
cially prescribed perfectionism entails the belief that others
have perfectionistic expectations and motives for oneself.
The need for a multidimensional approach is indicated by
research with college students that has shown that both self-or-
iented and socially prescribed perfectionism are correlated posi-
tively with subclinical depression (Flett, Hewitt, Blankstein, &
O'Brien, in press; Hewitt & Flett, in press). These findings not-
withstanding, it remains to be established whether the various
perfectionism dimensions are involved in clinical depression.
Thus, the primary goal o f this study was to use a multidimen-
sional perspective to examine whether clinically depressed per-
sons are characterized by high levels of perfectionism.
Although our main purpose was to examine perfectionism
and clinical depression, we felt it was also important to examine
perfectionism and other forms o fpsychopathology, such as anxi-
ety disorders. A link with anxiety is suggested by past research
that has shown an association between self-oriented perfection-
istic attitudes and trait anxiety in college students (Flett, He-
witt, & Dyck, 1989). An association between socially pre-
scribed perfectionism and anxiety is particularly likely given
that a fear of negative social evaluation is an important element
of both constructs (Endler & Okada, 1975), as well as the fact

98
 SHORT REPORTS
that the discrepancy between the actual self and the ought self
has been related to greater anxiety (Higgins, Bond, Klein, &
Strauman, 1986; Strauman, 1989). Consequently, a second goal
o f this study was to determine if differences in perfectionism
were specific to depression or generalize to anxiety. This was
addressed by c o m p a r i n g perfectionism levels in depressed pa-
tients, anxious patients, and n o r m a l control subjects.
Method
Subjects
The depressed group consisted of 22 patients (6 men and 16 women)
admitted to an acute-care psychiatric unit. These patients were inter-
viewed by a psychiatrist and were diagnosed as having unipolar de-
pression on the basis of criteria from the Diagnostic and Statistical
Manual of Mental Disorders (rev. 3rd ed.; DSM-III-R; American Psy-
chiatric Association, 1987). Each patient also had a probable or defi-
nite diagnosis of major depression on the basis of Research Diagnostic
Criteria (Spitzer, Endicott, & Robins, 1978) and Beck Depression In-
ventory scores (BDI; Beck, Rush, Shaw, & Emery, 1979) greater than 9.
Patients were excluded from the depressed group if they had a concur-
rent anxiety disorder. The depressed patients had a mean age of 35.64
years (SD = 11.38), mean years of education ofl 1.27 (SD = 2.39), and
mean number of previous admissions of 0.35 (SD = 0.59).
In addition, 13 patients (4 men and 9 women) diagnosed with an
anxiety disorder on the basis of DSM-III-R (American Psychiatric
Association, 1987) criteria (4 simple phobia, 4 generalized anxiety, 3
obsessive-compulsive, and 2 panic disorder) were recruited from the
acute unit. Their mean age was 32.15 years (SD = 9.42), mean years of
education, 11.54 (SD = 1.13), and mean number of previous admis-
sions, 0.08 (SD = 0.28). Patients with a concurrent diagnosis of depres-
sion were excluded.
Finally, 22 normal control subjects (6 men and 16 women) were
matched on age and gender with the depressed group. Subjects were
excluded if they had BDI scores above 8 or if they had had any psycho-
logical treatment in the previous 2 years. Their mean age was 36.68
years (SD = 12.36), and they had a mean of 14.23 (SD = 1.60) years of
education.
Materials
Multidimensional Perfectionism Scale. The MPS (Hewitt & Flett,
1989, in press) has three subscales of 15 items each. Respondents make
7-point ratings (1 = stronglydisagreeto 7 = stronglyagree) of statements
that reflect Self-Oriented Perfectionism (e.g., One of my goals is to be
perfect in every thing I do), Other-Oriented Perfectionism (e.g., I have
high expectations for the people who are important to me), and Socially
Prescribed Perfectionism (e.g., My family expects me to be perfect).
Coefficient alphas were reported as .88 for Self-Oriented, .74 for Other-
Oriented, and .81 for Socially Prescribed Perfectionism (Hewitt &
Flett, in press), and test-retest reliabilities over 3 months in 39 patients
were .75, .65, and .78 for these subscales, respectively. Intercorrelations
of the subscales range between .25 and .40 for students and between .28
and .53 for patients. Additional evidence indicates that the MPS sub-
scales have adequate concurrent validity in clinical samples (Hewitt &
Flett, in press; Hewitt, Flett, Turnbull-Donovan, & Mikail, 1990).
Beck Depression Inventory The BDI is a 2 l-item scale designed to
measure the severity of depressive symptomatology (Beck et al., 1979).
Several studies have demonstrated its reliability and validity (Beck,
Steer, & Garbin, 1988).
Endler Multidimensional Anxiety Scales-State (EMAS-S; Endler,
Edwards. Vitelli, & Parker, 1989). The EMAS-S is a 20-item self-re-
port measure of the autonomic-emotional and cognitive-worry compo-
99
nents of state anxiety. Extensive validity and reliability evidence has
been provided by Endler et al. (1989).
Procedure
Depressed and anxious patients with a minimum Grade 8 education
and no history of mania, psychosis, organic impairment, or alcohol
abuse were asked to volunteer in a study of personality factors and
depression. Patients were interviewed by Paul L. Hewitt to confirm the
DSM-III-R diagnoses from the Research Diagnostic Criteria (Spitzer
et al., 1978) and to assess inclusion and exclusion criteria. Initially,
three potential subjects, two with diagnoses of depression and one of
anxiety, chose not to participate in the study. The normal control sub-
jects were recruited from hospital staff and the community and were
selected initially on the basis of matches with the depressed patients.
These subjects were interviewed to assess inclusion and exclusion crite-
ria. Questionnaires were administered in a random order and all sub-
jects were paid $20 Canadian for their participation.
Results
Initially, several one-way analyses o f variance were done to
compare groups on age and education. The three groups did
not differ on age, F(2, 54) = 0.67, ns, but did differ on educa-
tion, F(2, 54) = 15.74, p < .001. The normal control subjects
had had m o r e e d u c a t i o n t h a n had either the depressed or
anxious patients. The patient groups did not differ from one
another. All o f the analyses herein used education level as a
covariate.
The means and standard deviations o f the measures are pre-
sented in Table 1. An analysis o f covariance with BDI scores as
the dependent variable was conducted to compare depression
scores a m o n g the groups. This analysis was significant, F(2,
53) = 30.73, p < .001, and multiple comparisons showed that
depressed patients had higher BDI scores than did either the
anxiety patients or the normal control subjects. In addition, the
anxiety patients had higher BDI scores than did the normal
control subjects. Anxiety levels were also c o m p a r e d a m o n g the
groups in a one-way analysis o f covariance. This analysis was
significant, F(2, 53) = 17.88, p < .001, wherein both the de-
pressed and anxious patients had higher E M A S - S scores than
did the normal control subjects but did not differ from one
another.
G r o u p differences on the perfectionism dimensions were ex-
a m i n e d by conducting a multivariate analysis o f covariance
with group status as the independent variable and the M P S
subscale scores as the dependent variables. The multivariate
effect o f group status was significant, F(6, 104) = 3.28, p < .01.
Univariate tests indicated that the three groups differed on Self-
Oriented Perfectionism, F(2, 53) = 3.87, p < .05, and on So-
cially Prescribed Perfectionism, F(2, 53) = 7.04, p < .001, but
not on Other-Oriented Perfectionism, F(2, 53) = 0.01, ns. Multi-
ple comparisons revealed that depressed patients had higher
m e a n levels o f Self-Oriented Perfectionism than did the other
two groups, but the anxiety patients and normal control sub-
jects did not differ from one another. Finally, depressed and
anxious patients had higher levels o f Socially Prescribed Perfec-
tionism than did the normal control subjects. The depressed
and anxious patients did not differ from one another on So-
cially Prescribed Perfectionism.
100 SHORT REPORTS

Table 1
Means and Standard Deviations of Depression, Anxiety, and Perfectionism Measures
for Depressed, Anxious, and Control Subjects
Depressed Anxious Control Total

Measure M SD M SD M SD M SD
BDI 25.32 9.59 18.76 11.18 2.96 0.63 15.12 12.98
EMAS-S 54.09 18.03 50.39 16.19 22.05 2.42 40.87 20.23
MPS
Self 76.05 17.46 64.39 15.61 63.50 15.70 68.54 17.17
Other 52.00 15.81 52.31 13.60 55.46 11.55 53.40 13.64
Social 60.50 20.10 58.39 13.11 42.64 12.68 53.12 17.88

Note. The data are based on the responses of 22 depressed patients, 13 anxious patients, and 22 control
subjects. Higher scores reflect greater depression, anxiety, and perfectionism. BDI = Beck Depression
Inventory; EMAS-S = Endler Multidimensional Anxiety Scales-State; MPS = Multidimensional Perfec-
tionism Scale; Self = Self-Oriented Perfectionism; Other = Other-Oriented Perfectionism; and Social =
Socially Prescribed Perfectionism.

Table 2 presents the correlations among the measures col- cally anxious and normal control subjects. The results showed
lapsed across groups. ~ It can be seen that both Self-Oriented that the depressed patients were differentiated from the other
and Socially Prescribed Perfectionism were associated signifi- subjects by a higher level of self-oriented perfectionism. Thus,
cantly with depression and anxiety. Regression analyses were our results suggest that higher levels of self-oriented perfection-
conducted to provide further information with regard to the ism may be specific to clinical depression and do not generalize
unique contribution of the MPS dimensions to depression and to clinical anxiety.
anxiety. In this set of analyses, after anxiety scores were first Self-oriented perfectionism ought to be related to depression
entered, Self-Oriented Perfectionism contributed a significant for several reasons. Self-oriented perfectionists' tendencies to
amount of variance in depression scores (multiple R = .82, set unrealistic standards and stringently evaluate their own per-
R2ehaage= .03, p < .05). Similarly, again after anxiety scores were formance increases not only the frequency of failure (Kanfer &
entered, Socially Prescribed Perfectionism accounted for addi- Hagerman, 1981) but also the personal impact and meaning of
tional variance (multiple R = .83, R2c~n~ = .04, p < .05). Other- failure experiences. Because self-oriented perfectionists tend to
Oriented Perfectionism did not contribute significant unique equate self-worth with performance (Pacht, 1984), falling short
variance. The analyses to predict anxiety found that none of the of self-imposed standards on a consistent basis may promote
MPS subscales contributed significant variance after depres- chronic deficits in self-esteem and self-evaluation. This sug-
sion had been entered into the equation. gests that these persons may generate their own failures and
stressors, which make them particularly prone to depressive
episodes.
Discussion Consistent with expectations, additional findings revealed
The purpose of this study was to use a multidimensional that both the depressed and anxious patients had higher levels
approach to assess whether levels of perfectionism in clinically of socially prescribed perfectionism than did the normal con-
depressed patients differ from levels of perfectionism in clini- trol subjects. These data indicate that socially prescribed per-
fectionism is a feature of depression, but it is not necessarily
specific to depression. This corroborates previous results with
college students that indicated that socially prescribed perfec-
Table 2 tionism is related to several types of maladjustment (Hewitt &
Correlations Among Perfectionism Dimensions, Flett, in press). It is also consistent with claims that a neurotic
Depression, and Anxiety form of anxious depression exists and stems in part from per-
ceived deprivation and lack of gratification from significant
Measure 1 2 3 4
others (see Gersh & Fowles, 1979; Kiloh & Garside, 1963).
MPS The analyses further revealed that other-oriented perfection-
1. Self ism was not related to depression in this study. Although this
2. Other .35** -- dimension was not associated with depression or with anxiety,
3. Social .49*** .24 --
4. BDI .42** .06 .61"** B
other-oriented perfectionism has been shown to correlate with
5. EMAS-S .32* -.11 .56*** .80"**

Note. MPS = Multidimensional Perfectionism Scale; Self = Self-Or-

iented Perfectionism; Other = Other-Oriented Perfectionism; and So- i These correlations and subsequent regression analyses must be in-
cial = Socially Prescribed Perfectionism; BDI = Beck Depression In- terpreted with caution, given that several assumptions are violated
ventory; EMAS-S = Endler Multidimensional Anxiety Scales-State. when treating data from discrete groups as continuous. We included
*p<.05. **p<.01. ***p<.001. these analyses because of the relative paucity of data in this area.
 SHORT REPORTS
measures of antisocial, histrionic, and narcissistic personality
disorders (see Hewitt & Flett, in press).
Certain limitations of the current study must be noted. First,
no attempt was made to examine factors that may mediate the
association between perfectionism and depression. Future re-
search needs to examine possible mediating factors such as life
stress (Hewitt & Dyck, 1986), self-critical attributional style,
self-focused attention, and maladaptive coping styles (Flett et
al., in press). A consideration of these factors will be in keeping
with self-regulation models (e.g., Kanfer & Hagerman, 1981),
which maintain that the manner in which perfectionists inter-
pret and cope with failure is an important determinant of their
experience of depression.
Second, it is apparent that our findings are based on self-re-
port data and ought to be replicated with behavioral measures.
Third, our results do not address causality or vulnerability is-
sues. Although the findings indicated that perfectionism is as-
sociated with clinical depression and anxiety during the depres-
sive or anxious episode, it remains for future research to estab-
lish whether the perfectionism dimensions are involved in
susceptibility to these adjustment problems. Finally, it is essen-
tial that future research determines whether the various dimen-
sions of perfectionism account for unique variance in depres-
sion over and above the variance accounted for by other person-
ality factors involved in depression.
In summary, this study compared levels of perfectionism in
depressed and anxious patients and normal control subjects.
Depressed patients were differentiated from the other two
groups by higher levels of self-oriented perfectionism. In addi-
tion, both depressed and anxious patients were characterized
by elevated levels of socially prescribed perfectionism+ These
findings support the usefulness of further research of the role of
perfectionism in clinical depression.
References
American Psychiatric Association. (1987). Diagnostic and statistical
manual of mental disorders (Rev. 3rd ed.). Washington, DC: Author.
Beck, A. T. (1967). Depression: Clinical experimental and theoretical
aspects. New York: Harper & Row.
Beck, A. T., Rush, A. J., Shaw, B. E, & Emery, G. (1979). Cognitive
therapy of depression: A treatment manual. New York: Guilford
Press.
Beck, A. T., Steer, R, A., & Garbin, M. G. (1988). Psychometric proper-
ties of the Beck Depression Inventory: Twenty-fiveyears of evalua-
tion. Clinical Psychology Review, 8, 77-100.
Bibring, E. (1953). The mechanism of depression. In E Greenacre
(Ed.), Affective disorders (pp. 13-48). New York: International Uni-
versities Press.
Burns, D. D. (1983). The spouse who is a perfectionist. Medical Aspects
of Human Sexuality, 17, 219-230.
Carson, R. C. (1989). Personality. Annual Review of Psychology, 40,
227-248.
Carver, C. S., LaVoie, L., Kuhl, J., & Ganellen, R. J. (1988). Cognitive
concomitants of depression: A further examination of the roles of
generalization, high standards, and self-criticism. Journal of Social
and Clinical Psychology, 7, 350-365.
Endler, N. S., Edwards, J. M., Vitelli, R., & Parker, J. D. A. (1989).
101
Assessment of state and trait anxiety: Endler Multidimensional Anx-
iety Scales. Anxiety Research, 2, 1-14.
Endler, N. S., & Okada, M. (1975). A multidimensionalmeasure o ftrait
anxiety: The S-R Inventory of General Trait Anxiousness. Journalof
Consulting and Clinical Psychology, 43, 319-329.
Flett, G. L., Hewitt, P. L., Blankstein, K. R., & O'Brien, S. (in press).
Perfectionism and learned resourcefulness in depression and self-es-
teem. Personality and Individual Differences.
Flett, G. L., Hewitt, P. L., & Dyck, D. G. (1989). Self-oriented perfec-
tionism, neuroticism, and anxiety.Personality and Individual Differ-
ences, 10, 731-735.
Frost, R., Marten, E, Lahart, C., & Rosenblate, R. (1990). The dimen-
sions of perfectionism. Cognitive Therapy and Research, 14, 449-
468.
Ganellen, R. J. (1988). Specificity of attributions and overgeneraliza-
tion in depression and anxiety. Journal of Abnormal Psychology, 97,
83-86.
Gersh, R., & Fowles, D. (1979). Neurotic depression: The concept of
anxious depression. In R. Depue (Ed.), The psychobiology of depres-
sive disorders: Implications for the effects of stress (pp. 81-104). San
Diego, CA: Academic Press.
Hewitt, P. L., & Dyck, D. G. (1986). Perfectionism, stress, and vulnera-
bility to depression. Cognitive Therapy and Research, I 0, 137-142.
Hewitt, P. L., & Flett, G. L. (1989). The Multidimensional Perfection-
ism Scale: Development and validation [Abstract]. Canadian Psy-
chology, 30, 339.
Hewitt, E L., & Flett, G. L. (1990). Perfectionism and depression: A
multidimensional analysis. Journal of Social Behavior and Personal-
ity 5, 423-438.
Hewitt, P. L., & Flett, G. L. (in press). Perfectionism in the self and
social contexts: Conceptualization, assessment, and association
with psychopathology. Journal of Personality and Social Psychology
Hewitt, P. L., Flett, G. L., Turnbull-Donovan, W.,& Mikail, S. E (1990).
The Multidimensional Perfectionism Scale: Reliability, validity, and
psychometric properties in psychiatric samples. Manuscript submit-
ted for publication,
Hewitt, E L., & Genest, M. (1990). The ideal-self: Schematic process-
ing of perfectionistic content in dysphoric university students. Jour-
nal of Personality and Social Psychology, 59, 802-808.
Higgins, E. T., Bond, R. N., Klein, R., & Strauman, T. (1986). Self-dis-
crepancies and emotional vulnerability: How magnitude, accessibil-
ity, and type of discrepancy influence affect. Journal of Personality
and Social Psychology, 51.5-15.
Kanfer, E H., & Hagerman, S. (1981). The role of self-regulation. In L.
Rehm (Ed.), Behavior therapy for depression: Present status and fu-
ture directions (pp. 143-179). San Diego, CA: Academic Press.
Kiloh, L. G., & Garside, R. E (1963). The independence of neurotic
depression and endogenous depression. British Journal of Psychia-
try, 109, 451-463.
Pacht, A. R. (1984). Reflectionson perfection. American Psychologist,
39, 386-390.
Spitzer, R. L., Endicott, J., & Robins, E. (1978). Research diagnostic
criteria: Rationale and reliability.Archives of General Psychiatry, 35,
773-782.
Strauman, T. J. (1989). Self-discrepancies in clinical depression and
social phobia: Cognitive structures that underlie emotional dis-
orders? Journal of Abnormal Psychology, 98, 14-22.
Received December 12, 1989
Revision received August 17, 1990
Accepted August 28, 1990 •