Behavior Modification

http://bmo.sagepub.com/

So Perfect It's Positively Harmful?: Reflections on the Adaptiveness

and Maladaptiveness of Positive and Negative Perfectionism
R. Glynn Owens and Peter D. Slade
Behav Modif 2008 32: 928 originally published online 3 June 2008
DOI: 10.1177/0145445508319667

The online version of this article can be found at:

http://bmo.sagepub.com/content/32/6/928

Published by:

http://www.sagepublications.com

Additional services and information for Behavior Modification can be found at:

Email Alerts: http://bmo.sagepub.com/cgi/alerts

Subscriptions: http://bmo.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

Citations: http://bmo.sagepub.com/content/32/6/928.refs.html

>> Version of Record - Oct 23, 2008

OnlineFirst Version of Record - Jun 3, 2008

What is This?

Downloaded from bmo.sagepub.com at UNIV TORONTO on November 14, 2014

 Behavior Modification
Volume 32 Number 6
November 2008 928-937
© 2008 Sage Publications
So Perfect It’s Positively 10.1177/0145445508319667
http://bmo.sagepub.com
Harmful? hosted at
http://online.sagepub.com

Reflections on the Adaptiveness and

Maladaptiveness of Positive and
Negative Perfectionism
R. Glynn Owens
University of Auckland
Peter D. Slade
Formerly University of Liverpool

The article by Flett and Hewitt (2006) highlights a number of important issues
in the study of perfectionism and rightly urges caution against simplistic con-
ceptualizations. Their view that the term perfectionism should be reserved for
pathological forms of behavior is questionable, though understandable given
the perspective from which they view it. But whatever the terminology used,
the underlying processes remain unaffected. Relevance of the data they cite in
support of an alleged maladaptive side to positive perfectionism, however,
rests on the assumption of close parallels between self-oriented perfectionism
and positive perfectionism, an assumption that may not be justified. It is
important to draw attention not only to differences between the theoretical
underpinnings of their and the authors perspectives but also to points of agree-
ment that the authors may previously have failed to make clear. The authors
concur entirely with their view that clear avenues of potential research are now
apparent that should serve to clarify the issues.

Keywords: positive perfectionism; negative perfectionism; Multidimensional

Perfectionism Scale; self-oriented perfectionism; reinforcement

F lett and Hewitt’s (2006) article is generous in its praise for our attempts to
clarify the processes underlying perfectionistic behavior, and their com-
ment that it “is an important contribution to this field” is gratifying. In dis-
cussing the model we have proposed, however, they raise a number of valuable
and important concerns. These may be briefly summarized as follows:

Authors’ Note: Please address correspondence to R. Glynn Owens, Department of Psychology

(Tamaki) Private Bag, 92019, Auckland, New Zealand; e-mail: g.owens@auckland.ac.nz.

928

Downloaded from bmo.sagepub.com at UNIV TORONTO on November 14, 2014

 Owens, Slade / Reflections on the Adaptiveness and Maladaptiveness 929

1. What we have termed “positive perfectionism” should not be considered

a form of perfectionism, and a different term like “conscientiousness”
should be used instead.
2. Their own considerable research has shown that their dimension of self-
oriented perfectionism is by no means unequivocally positive, and
indeed may have many negative and maladaptive features.
3. They question the extent to which positive and negative perfectionism
are adequately distinguished and suggest that there may be considerable
correlation between the two.

The considerable evidence for these objections is presented cogently and

clearly in their article. Nevertheless, we remain unconvinced by the argu-
ments presented. We believe that the concerns raised are a reflection, at
least in part, of our failure to make clear certain points in our earlier expo-
sitions. These are discussed in somewhat more detail below with respect to
each of the three concerns mentioned. It is perhaps helpful if we first, however,
provide a brief summary of the model we have developed (Slade & Owens,
1998), clarifying some issues where we feel we may have previously caused
some confusion.
Although the suggestion of a positive form of perfectionism is sometimes
seen as the most radical element of this model, a more fundamental aspect
may be the shift from viewing perfectionism as a characteristic of the indi-
vidual to viewing it as a characteristic of behavior. Put simply, we suggest
that more important than the form of the behavior are the consequences
attendant on it. In this respect, our perspective is remarkably close to that of
Flett and Hewitt (2006) when they remark that “conclusive statements about
whether perfectionism is positive or negative cannot be made without taking
into account the outcomes that the perfectionist is experiencing in his or her
environment” (p. 479). Our own position is that whether perfectionistic
behavior is positive or negative is determined by whether the relevant rein-
forcers of the behavior are positive or negative. Of course, this is not to deny
that characteristics of the individual are important; at the most basic level,
whether a particular consequence is reinforcing or not will vary across indi-
viduals, as will whether such reinforcers are positive or negative.
Seen in this light, a crucial difference between the positive and negative
perfectionist is likely to be the emotional connotations of the perfectionistic
behavior. When our behavior is positively reinforced, we tend to regard it as
behavior we “choose” to do, behavior that is under our control—acting “of
our own free will” it could be said. By contrast, under conditions of negative
reinforcement, we tend to feel that we are “coerced” and “have no control”
or “no choice” in our actions. In this respect, one might regard positive

Downloaded from bmo.sagepub.com at UNIV TORONTO on November 14, 2014

930 Behavior Modification

perfectionism as perfectionism we choose and negative perfectionism as per-

fectionism that is forced upon us. Of course, to the outside observer, the actual
behaviors may look identical—the rat that presses the lever to obtain food
looks indistinguishable from the one that presses the lever to avoid shock.
We are thus suggesting that one useful perspective in the study of per-
fectionism may be to consider it in terms of a style of behavior and the ways
in which such behavior is reinforced. Of course, any particular instance of
behavior may produce several reinforcers, conceivably a mixture of both
positive and negative ones. Moreover, it is clear that contingencies main-
taining such a style of behavior may change over time, again possibly from
the positive to the negative; indeed, our original article alludes to this pos-
sibility directly, giving the example of the athlete whose positive reinforce-
ment associated with early success is later subject to negative reinforcement
as a result of the expectations produced. Adopting this kind of behavioral
model in the study of perfectionism means something of a conceptual jump
in the way it is conceived, and it is not surprising that from time to time
misunderstandings and misinterpretations occur. We suspect that much of
the concern raised in Flett and Hewitt’s (2006) article results more from
such misunderstanding than from specific problems with the model.

Can Perfectionism Be Positive?

As Flett and Hewitt point out, they approach the issue of perfectionism
mainly from the viewpoint of having been concerned with its clinical mani-
festations. In this respect, it is unsurprising that they are reluctant to concede
the possibility of a positive aspect, having seen at first hand (as we have too)
the suffering that may be associated with severely destructive negative forms
of perfectionism. Correspondingly, it is perhaps similarly unsurprising that
our own work, spanning both clinical cases and groups such as sportspeople,
should have exposed us to both positive and negative forms and permitted us
to consider both possibilities. In this sense, it is understandable that they feel
uncomfortable with the use of a term like “perfectionism” to describe some-
thing that is largely positive. But like Shakespeare’s Juliet, we may ask
“What’s in a name?” If we by edict declare that what we call “positive per-
fectionism” should henceforth be known as “positive striving” or whatever,
this will affect not one bit the relevant psychological processes that have
been going on long before we decided to give them names. Given, however,
that the notion of positive perfectionism is (as Flett & Hewitt, 2006, point
out) already in common usage, and that many of the accepted measures of

Downloaded from bmo.sagepub.com at UNIV TORONTO on November 14, 2014

 Owens, Slade / Reflections on the Adaptiveness and Maladaptiveness 931

perfectionism appear to tap into the dimension we describe, it is better that

we should break the mind-set that forces us to see perfectionism as inevitably
negative rather than expect the scientific and lay community to change the
terms they use to describe behavior. Such a change might be arguable if it
could be demonstrated that the term was a clear misnomer. Consideration of
such a change might, therefore, be best made once we have considered the
second of Flett and Hewitt’s concerns, that of whether the evidence supports
a positive form.

Can Positive Perfectionism Be Maladaptive?

Flett and Hewitt (2006) cite an impressive body of evidence suggesting

possible problems with a notion that positive perfectionism is necessarily
adaptive. The overwhelming bulk of the evidence cited depends on their
assumption that “self-oriented perfectionists . . . would be characterized by
Slade and Owens . . . as having a healthy or positive form of perfectionism”
(p. 479). It seems likely that Flett and Hewitt base this assumption on Table
1 (p. 377) in our article, and the accompanying text, where we assign self-
oriented perfectionism to the Type 1 category and socially prescribed per-
fectionism to the Type 2 category. This was based on the study of Frost,
Heimburg, Holt, Mattia, and Neubauer (1993), which found that self-oriented
perfectionism loaded on their factor of “positive striving,” whereas socially
prescribed perfectionism loaded on their factor of maladaptive evaluation
concerns. However, we believe that self-oriented perfectionism (unlike pos-
itive perfectionism) may be associated with both adaptive and maladaptive
behavior for both empirical and theoretical reasons. We apologize for any
confusion in this respect; it has never been our claim that self-oriented
perfectionism is synonymous with positive perfectionism, and indeed it is
slightly surprising that Flett and Hewitt should have made this mistake,
given that they quote (p. 476) sample items from our scale which they them-
selves remark are associated with social approval (and therefore not, pre-
sumably, self-oriented perfectionism). It also seems somewhat inconsistent
that in one part of their article, they should argue that the term perfectionism
should not be used for what we have termed positive perfectionism and then
to assume that this same dimension is measured by their own perfectionism
scale. In fact, close examination of the issues reveals both empirical and con-
ceptual reasons why it would be a serious error to assume any simple link
between positive and self-oriented perfectionism.
The empirical reasons primarily concern the way in which “self-oriented”
perfectionism is measured. The main tool is of course Hewitt and Flett’s

Downloaded from bmo.sagepub.com at UNIV TORONTO on November 14, 2014

932 Behavior Modification

(1991) Multidimensional Perfectionism Scale (MPS). This contains three

separate scales, Self-Oriented, Socially Prescribed, and Other-Oriented
Perfectionism. However, the items measuring these scales have a mixed
format. Some are written in such a way as to reflect a mixture of “positive
reinforcement” and “negative reinforcement” with others ambiguous in this
regard. Thus, many of the self-oriented perfectionism items are in fact also
negative perfectionism items. It follows, therefore, that in some data sets,
the negative items will have a predominant influence (greater than the pos-
itive and ambivalent items) leading to the empirical findings reviewed by
Flett and Hewitt in their article.
We have developed an alternative questionnaire measure of perfection-
ism, the Positive and Negative Perfectionism Scale (PANPS; Terry-Short,
Owens, Slade, & Dewey, 1995). This comprises 40 items; of which, 10 are
written in such a way as to measure self-oriented and positive perfectionism;
10 to measure self-oriented and negative perfectionism; 10 to measure
socially prescribed and positive perfectionism; and 10 to measure socially
prescribed and negative perfectionism. With this balanced format, we have
found that the positive–negative distinction overrides that between self-
oriented and socially prescribed perfectionism (Terry-Short et al., 1995).
Moreover, in another study by one of us (Celebi, Elal, & Slade, 1999), it has
been found that negative perfectionism is strongly correlated with both anx-
iety and depression, whereas positive perfectionism is not, with both self-
oriented and socially prescribed perfectionism exhibiting intermediate but
significant correlations. Moreover, in a subsequent analysis in which both
self-oriented and socially prescribed perfectionism were separated into sep-
arate 10 item positive and negative scales, the data show very clearly that it
is only the negative subscales (negative self-oriented and negative socially
prescribed perfectionism scales) that correlate significantly with anxiety and
depression. It seems, therefore, that the observed correlations of self-
oriented perfectionism with maladaptive functions, reviewed by Flett and
Hewitt (2006), are probably largely because of the fact that their MPS scales
include negative as well as positive items. Seen in this light, the correlations
they describe are neither surprising nor relevant to consideration of whether
positive perfectionism may have negative elements. It should be noted, inci-
dentally, that this argument holds good even if measures of positive perfec-
tionism and self-oriented perfectionism are shown to correlate, and it is
worth remembering that even if the two correlated .8, a correlation of .6
between self-oriented perfectionism and some negative characteristic would
still not necessarily imply any variance in common between that character-
istic and positive perfectionism. The correlations cited in their article are, of

Downloaded from bmo.sagepub.com at UNIV TORONTO on November 14, 2014

 Owens, Slade / Reflections on the Adaptiveness and Maladaptiveness 933

course, considerably lower than these, and no information is given about the
correlation between positive perfectionism and the measure they use.
One approach to resolving this issue would be if Flett and Hewitt were to
carry out further analyses (reanalyses) of their MPS data sets, with their
“Self-Oriented” Scale subdivided into separate “positive” and “negative”
subscales. We believe that this would take us someway toward understanding
the differences between their data and our own. An alternative would be the
simple step of using the PANPS (Terry-Short et al., 1995) to measure posi-
tive perfectionism directly rather than measures that are only assumed to have
some unquantified relationship. Certainly, it is clear that researchers who
have used specific measures of positive perfectionism have obtained results
that do not reflect at all the concerns raised by Flett and Hewitt (2006). Thus,
Mitchelson and Burns (1998) showed positive perfectionism to correlate with
life satisfaction and positive self-assessment, and Bergman, Nyland, and
Burns (2007), also using the scale we developed, have found no correlation
between positive perfectionism and a range of negative measures, including
regret, dysfunctional attitudes, and depression, all of which correlated signif-
icantly with negative perfectionism. Results such as these suggest the need
for careful matching of measures to hypotheses if appropriate conclusions are
to be drawn.
As mentioned earlier, there are also conceptual reasons to question Flett
and Hewitt’s (2006) conclusions. There seems, of course, no obvious reason
to assume that perfectionism directed toward oneself need necessarily be
associated with positive, rather than negative reinforcers, any more than
there is to assume that socially prescribed perfectionism will not be associ-
ated at times with positive reinforcement from others. For many individuals,
there may be a strong need to act in a self-oriented perfectionistic manner to
avoid some feared consequence. Such individuals would show considerable
self-oriented perfectionism, but this would fit within our concept of nega-
tive, not positive, perfectionism. Thus, although Flett and Hewitt cite a large
number of studies to indicate possible problems associated with self-oriented
perfectionism, there is as yet no obvious reason to conclude that these same
problems will necessarily be associated with positive perfectionism. Indeed,
Flett and Hewitt point out themselves that what they term “self-oriented
perfectionism” is not the same as positive perfectionism when they remark
“self-oriented perfectionism is associated with fear of failure and an intoler-
ance to failure” (p. 12, lines 18-19)—exactly the kind of characteristics asso-
ciated with negative perfectionism in our model.
It is perhaps important to note here that the issue is complex. Even if it
were the case that self-oriented perfectionism were more commonly subject

Downloaded from bmo.sagepub.com at UNIV TORONTO on November 14, 2014

934 Behavior Modification

to positive reinforcement, it seems likely that there would remain a sub-

stantial amount that was negatively reinforced, and such elements of nega-
tive reinforcement could account for many of the problems cited.
Having said which, there is no doubt that in some cases, Flett and Hewitt
will undoubtedly be correct. Pursuit of any positive reinforcer may under
some circumstances lead to problems. Most obvious is where a particular
activity constitutes almost the total source of reinforcement for that person.
Should the behavior for some reason cease to produce reinforcement, the
results may be problematic. This can be seen in a range of instances from
the observation of aggressive behavior with the onset of extinction in labo-
ratory animals (e.g., Skinner, 1968) or the development of depression in
humans (Ferster, 1973). However, in this context, it is important to note that
the problem is not one of perfectionism. Any individual who has only a lim-
ited number of sources of reinforcement, or a limited repertoire of behav-
iors to obtain reinforcement, is at risk should these later fail. But the same
problem would apply if the person’s reinforcement was almost exclusively
from a particular individual (e.g., a spouse who then leaves or dies) or a par-
ticular behavior (e.g., the person who obtains the overwhelming majority of
their reinforcement from playing a particular sport who is then forced to
retire through injury). The problem is not the perfectionism but “putting all
of one’s eggs in one basket” as far as sources of reinforcement are con-
cerned. In this sense, then, the exclusive pursuit of positive perfectionism
may indeed be maladaptive, but it is the exclusivity, not the perfectionism,
that is the problem.
Of course, there are also other mechanisms whereby positive perfec-
tionism may lead to undesirable consequences, some of which were out-
lined in our original article. For example, we have mentioned earlier that
what starts as positively perfectionistic behavior may become negatively
perfectionistic behavior through a change in reinforcement contingencies—
for example, when an athlete who gains pleasure from outstanding perfor-
mances begins to notice social pressure from team mates to succeed or the
individual who enjoys high-academic grades who then comes to interpret
any subsequent lower grade as an indication of failure.
Thus, although Flett and Hewitt’s (2006) article provides strong evi-
dence for the potentially maladaptive nature of self-oriented perfectionism,
the relevance to positive perfectionism remains unclear because none of the
studies they cite as showing significant correlations with negative factors
actually measures positive perfectionism. Nevertheless, they are, of course,
correct in indicating that even positive perfectionism should not be regarded
as an unqualified blessing or a panacea for all ills.

Downloaded from bmo.sagepub.com at UNIV TORONTO on November 14, 2014

 Owens, Slade / Reflections on the Adaptiveness and Maladaptiveness 935

The Discriminability of Positive

and Negative Reinforcement

Flett and Hewitt (2006), in their article, express concern that the dimen-
sions of positive and negative perfectionism may show some degree of over-
lap. They point out, quite correctly, that it is not unknown for scores on the
two dimensions to show some degree of correlation. Although this is not in
dispute, it is important here to distinguish the notions of scientific and statis-
tical independence. It will be remembered that the model here presented is
one of perfectionistic behavior maintained either by positive reinforcers, neg-
ative reinforcers, or both. Clearly, there is no problem in distinguishing the
processes of positive and negative reinforcement, and such processes are
demonstrated daily in operant laboratories around the world. In this sense, the
two concepts of positive and negative perfectionism show a straightforward
and easily understandable independence. Statistically, however, it would be
surprising if such independence were apparent when behavior such as per-
fectionistic behavior is evaluated in society at large. The world in which we
live provides a host of sources of reinforcement for perfectionistic behavior,
and it would be odd if individuals who exhibited such behavior only ever
received one or other form of reinforcement. Rather, we should expect that
when the behavior occurs it will not infrequently be associated with both pos-
itive and negative reinforcement. One implication of this would be that while
particular individuals might receive predominantly negative or positive rein-
forcers, there would be no obvious reason to assume that many would receive
only one type, although the underlying processes are distinct, statistically the
two would be correlated in society at large. This does not seem to us to be in
any sense a problem, and indeed it is notable that the correlations to which
Flett and Hewitt (2006) refer suggest that at best (as in the Lynd-Stevenson
& Hearne, 1999, study) less than half the variance is common between the
measures used, and when specific rather than proxy measures of positive and
perfectionism are used (as in the Mitchelson & Burns, 1998, study), only a
little more than 12% of the variance is common to the two. Such overlap does
not appear to be a basis for major concern.
Flett and Hewitt (2006) also refer to the possible problem of individuals
high in both positive and negative perfectionism experiencing consequent
approach-avoidance conflicts. At a simple level, of course, no conflict need
be apparent; conventionally, the term approach-avoidance conflict is used
when a particular course of action leads to both positive and negative
consequences—as with the rat in the box whose lever pressing produces
both food and shock. In the present instance, the parallel fails because the

Downloaded from bmo.sagepub.com at UNIV TORONTO on November 14, 2014

936 Behavior Modification

perfectionistic behavior is reinforced on both counts, positively and nega-

tively. A more direct parallel would be with the rat in the box whose behavior
both produces food and switches off an aversive stimulus, and no conflict
is apparent. It would be unfair, however, to object to this concern simply on
this basis. It is perfectly reasonable to expect that a number of individuals
whose perfectionistic behavior is both positively and negatively reinforced
may lack confidence in their ability to achieve a satisfactory level of perfor-
mance. In this sense, there may well be a dread of failure which would inhibit
striving for perfection and which would lead to exactly the type of approach-
avoidance conflict referred to. Once again, however, the issues are not ones
of the possible maladaptiveness of the positive perfectionism, or of the
correlation (which in any event appears weak) with negative perfectionism,
but rather the operation of negative reinforcers (and indeed punishers).

Conclusion

The article by Flett and Hewitt (2006) raises many relevant issues with
respect to our conceptualization of perfectionism and highlights the fact
that even with the most positive of characteristics it may still be valuable to
strike a cautionary note. Given, however, that much of their argument rests
on the assumption that self-oriented perfectionism (and some other dimen-
sions) are synonymous with positive perfectionism, we are not yet swayed
by the force of the arguments presented. Indeed, we suspect that much of
the apparent disagreement masks considerable commonality, and that in
some respects, the dispute may be no more than terminological. On one
aspect, beyond all else, we can certainly agree; forthcoming research in the
area promises to be exciting and valuable, and must be allowed to deter-
mine the eventual outcome of this debate.

References
Bergman, A. J., Nyland, J. E., & Burns, L. R. (2007). Correlates with perfectionism and the
utility of a dual process model. Personality and Individual Differences, 43, 389-399.
Celebi, E., Elal, G., & Slade, P. (1999). Perfectionism: healthy and pathological correlates.
Unpublished master’s thesis, Bogazici University, Istanbul, Turkey.
Ferster, C. B. (1973). A functional analysis of depression. American Psychologist, 28, 857-870.
Flett, G. L., & Hewitt, P. L. (2006). Positive versus negative perfectionism in psychopathology:
A comment on Slade and Owens’s dual process model. Behavior Modification, 30, 472-495.
Frost, R. O., Heimburg, R. G., Holt, C. S., Mattia, J. L., & Neubauer, A. L. (1993). A compari-
son of two measures of perfectionism. Personality and Individual Differences, 4, 119-126.

Downloaded from bmo.sagepub.com at UNIV TORONTO on November 14, 2014

 Owens, Slade / Reflections on the Adaptiveness and Maladaptiveness 937

Hewitt, P. L., & Flett, G. L. (1991). Perfectionism in the self and social contexts:
Conceptualization, assessment and association with psychopathology. Journal of Personality
and Social Psychology, 60, 456-470.
Lynd-Stevenson, R. M., & Hearne, C. (1999) Perfectionism and depressive affect: The pros
and cons of being a perfectionist. Personality and Individual Differences, 26, 549-562.
Mitchelson, J. K., & Burns, L. R. (1998). Career mothers and perfectionism: Stress at work
and stress at home. Personality and Individual Differences, 25, 477-485.
Skinner, B. F. (1968). Contingencies of reinforcement: A theoretical analysis. Norwalk, CT:
Appleton-Century-Crofts.
Slade, P. D., & Owens, R. G. (1998). A dual process model of perfectionism based on rein-
forcement theory. Behavior Modification, 22, 372-390.
Terry-Short, L. A., Owens, R. G., Slade, P. D., & Dewey, M. E. (1995). Positive and negative
perfectionism. Personality and Individual Differences, 18, 663-668.

R. Glynn Owens received a BTech (Hons) in psychology from Brunel University, England, in
1974, a DPhil from the University of Oxford in 1977, and the diploma in clinical psychology
from the British Psychological Society in 1979. He was a lecturer and senior lecturer at the
University of Liverpool where he was director of the clinical psychology training program
before being appointed as the first professor of health studies at the University of Wales in 1992,
professor of psychology at the University of Auckland, New Zealand, in 1995, and professor of
forensic clinical psychology at the University of Wales in 2003. He returned to his post at the
University of Auckland in 2005. He is the author of more than 100 psychological publications
on topics including health psychology, clinical psychology, and forensic psychology. He is the
current president of New Zealand’s Independent Forensic Practitioners Institute.

Peter D. Slade received a BA (Hons) in psychology from the University of Sheffield in 1965,
MPhil in clinical psychology from the University of London Institute of Psychiatry in 1969,
and a PhD from the University of London in 1975. In 1990, he was appointed as a personal
chair in clinical psychology at the University of Liverpool, a position from which he retired in
1996. He is the author of more than 140 psychological publications on topics including eating
disorders, health psychology, and the psychology of psychotic behavior.

Downloaded from bmo.sagepub.com at UNIV TORONTO on November 14, 2014