Professional Documents
Culture Documents
Abstract
Nonsuicidal self-injury (NSSI) is quite common in eating disorder (ED) patients and we wondered whether this combined self-harming
behaviour is related to perfectionism, a feature often found in ED patients. In addition, we examined associations between perfectionism
and functions underlying NSSI and the possible mediating role of intrapersonal perfectionism in the association between perceived parental criticism and NSSI. In a sample of 95 ED patients, 38.9% reported at least one type of NSSI, and this subgroup reported signicantly higher levels of parental criticism and evaluative concerns perfectionism (ECP) compared with ED patients without NSSI. ECP was
positively related to the self-punishment and cry-for-help functions of NSSI. Finally, ECP was found to mediate the association between
parental criticism and NSSI symptoms. Directions for future research and practical implications are discussed. Copyright 2011 John
Wiley & Sons, Ltd and Eating Disorders Association.
Keywords
perfectionism; parental criticism; eating disorders; nonsuicidal self-injury
*Correspondence
Laurence Claes, Department of Psychology, Catholic University of Leuven, Tiensestraat 102, 3000 Leuven, Belgium.
E-mail: Laurence.claes@psy.kuleuven.be
L. Claes et al.
L. Claes et al.
Method
Participants
The ED sample consisted of 95 women admitted to a specialised
inpatient treatment unit. Patients were diagnosed according to
Diagnostic and Statistical Manual of Mental Disorders-IV
criteria (American Psychiatric Association, 1994) on the basis
of a standardised interview and questionnaire (Eating Disorder
Evaluation Scale; Vandereycken, 1993): 44 patients were diagnosed as anorexia nervosa, restrictive type, 12 as anorexia
nervosa, bingeing-purging type, 28 as bulimia nervosa and 11
as eating disorder not otherwise specied. The age of the ED
sample ranged from 14 to 42 years, with a mean of 21.5 years
(SD = 6.23). No signicant age differences were found between
the different ED groups [F(3, 91) = 0.63, ns]. On the Self-Injury
Questionnaire (see further on), 37 patients (38.9%) reported
at least one type of NSSI, whereas 58 (61.1 %) patients did
not display any type of NSSI. The presence/absence of NSSI
was not signicantly related to the different ED subtypes
(w2(3) = 5.07, ns). Of the 37 NSSI patients, 20 (21.1%) displayed
cutting, 19 (20.0%) hair pulling, 14 scratching (14.7%), 12
bruising (12.6%) and four (4.2%) burning.
Procedure
Eating disorder patients completed questionnaires as a part of the
routine assessment at admission. Participation was voluntary and
anonymity was guaranteed. All participants gave written informed
consent. The study was approved by the university Institutional
Review Board and by the Ethical Committee of the hospital
concerned. All items were rated on Likert scales ranging between
1 (completely disagree) and 5 (completely agree), unless indicated
otherwise.
Measures
Nonsuicidal self-injury. Using the Self-Injury Questionnaire (Claes
et al., 2003), patients were asked if they have deliberately injured
themselves (yes/no) in the past year by means of hair pulling,
scratching, bruising, cutting and/or burning. If so, they were to
specify how often this happened and which functions were served
by their NSSI (I feel the need to punish myself; I can physically
torture myself, to punish myself; I injure myself to show others
that something is wrong). Next, they were asked if they felt some
pain, and what kind of emotional experiences they had at the moment of self-injury. Additionally, patients were asked to provide
information about the age of onset of NSSI and the body parts
that were injured.
Intrapersonal perfectionism. The participants completed the
Dutch version (Soenens, Luyckx et al., 2008; Soenens, Vansteenkiste
et al., 2008) of the Multidimensional Perfectionism Scale (MPS;
Frost et al., 1990). In this article, we report on the ndings obtained
Eur. Eat. Disorders Rev. (2011) 2011 John Wiley & Sons, Ltd and Eating Disorders Association.
with the three subscales of the MPS reecting intrapersonal perfectionism: CM (nine items, I should be upset if I make a mistake),
doubts about actions (DA, four items, I usually have doubts about
the simple everyday things I do), and PS (seven items, I set higher
goals for myself than most people). Previous factor-analytical studies have shown that items from the CM and the DA subscales load
together on a single factor that has been labelled maladaptive perfectionism or, as we prefer, ECP, whereas the items of the PS load
together on a separate factor labelled adaptive perfectionism or
PSP (Dunkley, Blankstein et al., 2006; Dunkley, Sanislowet al.,
2006). To examine the distinction between ECP and PSP in our
sample, we performed a Principal Components Analysis of the 20
items for intrapersonal perfectionism. Two components had an
eigenvalue larger than 2 and the scree-plot also pointed to a twocomponent solution, explaining 48.93 % of the variance. The rst
component was mainly dened by items from the CM and DA subscales. The second component was dened by items from the PS
subscale. Given these ndings, we computed an average score of
the items for CM and DA, and we used this score as a measure of
ECP (Cronbachs a = .84). The score for PS was used as a measure
of PSP (Cronbachs a = .88; for this procedure, see Dunkley,
Blankstein, et al., 2006; Dunkley, Sanislow et al., 2006; Soenens,
Luyckx, et al., 2008; Soenens, Vansteenkiste, et al., 2008).
Perceived parental criticism. To assess PPC, we used another
subscale from the MPS, the Parental Criticism Scale (four items,
As a child, I was punished for doing things less than perfectly).
This scale has been used as a measure of parental criticism in previous research on parenting, perfectionism and NSSI (Yates et al.,
2008). Cronbachs a in this study was .72.
Severity of eating disorder symptoms. Because we aimed to
control for severity of ED symptoms when predicting NSSI, the
participants completed the three main scales from the Dutch
version (Van Strien, 2002) of the Eating Disorder Inventory-II
(EDI-II) (Garner, 1991), that is, drive for thinness (Cronbachs
a = .87), bulimia (Cronbachs a = .94), and body dissatisfaction
(Cronbachs a = .95). The participants rated how much each item
applied to them on a scale ranging from 1 (never) to 6 (always).
Information on the validity and psychometric characteristics of
the EDI-II is provided in Garner (1991).
Statistical analyses
The examine the mean level differences in ECP, PSP and PPC
between ED patients with and without NSSI, we performed a
multivariate ANCOVA with the presence/absence of NSSI as independent variable, ED severity as covariate and the perfectionism
scales as dependent variables. To investigate the associations between the perfectionism scales and the possible functions of NSSI,
we performed linear regression analyses with the functions of
NSSI as dependent variables and the perfectionism scales as independent variables or predictors.
And nally, to examine the mediating role of intrapersonal perfectionism between PPC and NSSI, we followed the procedure of Kenny,
Kashy and Bolger (1998). Step 1 involves determining the magnitude
of the association between the independent (i.e., PPC) and dependent
variable (i.e., presence/absence of NSSI). Step 2 requires nding a signicant association between the independent (i.e., PPC) and mediating variable (i.e., intrapersonal perfectionism). Step 3 requires nding
a signicant association between the mediating and dependent
L. Claes et al.
Results
Mean-level differences between patients with and
without nonsuicidal self-injury
Table 1 presents the means and standard deviations of ED patients
with and without NSSI on ECP, PSP and PPC controlled for ED
severity. A multivariate ANCOVA-analysis indicated that, across
all outcomes, both groups signicantly differed from each other
(Wilkss Lambda = 0.88, F(3, 83) = 3.65, p < .01). Subsequent univariate ANOVA-analyses indicated that after controlling for ED
severity, ED patients with NSSI scored signicantly higher on
ECP and PPC compared with ED patients without NSSI. However, both ED groups did not differ on PSP.
Associations between perfectionism and functions
of nonsuicidal self-injury
To investigate the association between the perfectionism scales and
possible functions of NSSI, we performed linear regression analyses
with the functions of NSSI (self-punishment, self-torturing, cry for
help) as dependent variables and ECP, PSP and PPC as independent
variables (Table 2). The results showed signicantly positive
associations between self-punishment/self-torturing and ECP. The
cry-for-help function showed a signicantly positive relation with
ECP and a negative relation with PPC.
Perfectionism as mediator between parental
criticism and nonsuicidal self-injury
To investigate the mediating role of intrapersonal perfectionism
between PPC and NSSI (see Figure 1), we followed the procedure
of Kenny et al. (1998). Because we did not nd an initial association between PSP and NSSI and because PPC was unrelated to
Table 1 Means and standard deviations of evaluative concerns perfectionism,
personal standard perfectionism and parental criticism for eating disorder patients
with and without nonsuicidal self-injury controlled for eating disorder severity
Evaluative concerns
Personal standards
Parental criticism
No NSSI
NSSI
(N = 55)
(N = 35)
SD
SD
F-value
3.55
3.78
2.10
0.75
0.78
0.78
3.92
3.76
2.60
0.68
0.86
1.09
5.58*
0.11
5.62*
Evaluative concerns
Personal standards
Parental criticism
Function 1
Function 2
Function 3
Self-punishment
Self-torturing
Cry-for-help
0.30*
0.04
0.13
0.30*
0.12
0.03
0.36**
0.01
0.29**
*p < 0.05
**p < 0.01, ***p < 0.001
Discussion
Although it is conceptually plausible to consider perfectionism as
a risk factor for NSSI, relatively few studies have empirically
addressed associations between perfectionism and NSSI. Moreover, most studies to date have relied on community samples.
The main aim of this study was to examine this association in a
sample where both perfectionism and NSSI are known to be elevated and salient. Indeed, we found that 38.9% of the ED inpatients in this study showed at least one type of NSSI, which is in line
with ndings of Svirko and Hawton (2007) who reported prevalence rates ranging from 25.4% to 55.2% in ED patients.
In examining associations between perfectionism and NSSI in
ED patients, we adopted, in line with the heavily growing body
of work on perfectionism outside the clinical domain, a twodimensional approach to perfectionism (Stoeber & Otto, 2006),
distinguishing between PSP (i.e., the setting of high standards
and goals for oneself) and ECP (i.e., the tendency to be concerned
about making mistakes and to engage in self-criticism). We found
that only ECP (but not PSP) was elevated among ED patients with
NSSI, even after controlling for severity of ED symptoms. This
Eur. Eat. Disorders Rev. (2011) 2011 John Wiley & Sons, Ltd and Eating Disorders Association.
L. Claes et al.
Perceived Parental
NSSI (0/1)
Criticism (PPC)
= 0.34***
Evaluative Concerns
Perfectionism (ECP)
=0.29**
( =0.22*)
Figure 1. The mediating role of intrapersonal perfectionism between perceived parental criticism and nonsuicidal self-injury. Beta coefcients without brackets refer
to the correlation coefcients between the variables that are linked by an arrow. Beta coefcients between brackets () refer to the regression weights of the regression analysis in which perceived parental criticism and evaluative concerns perfectionism were simultaneously included as predictors of nonsuicidal self-injury
L. Claes et al.
REFERENCES
scapegoating parents. First, we have assessed the patients perception of parenting and this (retrospective) view may be coloured by
the patients present mood and situation. Next, the causal direction of the relationship between parent behaviour and adolescent
psychopathology cannot be tested in a cross-sectional research,
but has to rely on a longitudinal design (see also Soenens, Luyckx,
et al., 2008; Soenens, Vansteenkiste, et al., 2008). Further, the
study could be replicated in a larger sample of ED patients (including also outpatients) to investigate whether the same associations
between parenting, perfectionism and NSSI remain in the different
diagnostic subtypes of ED. And nally, in future studies, we need
to control for Axes I and II comorbidity (e.g., borderline PD), impulsiveness and mood variability, given that these factors can also
inuence the association between PPC, perfectionism and NSSI.
Practical implications
The results of our study clearly indicated that ED patients with
NSSI show higher levels of ECP (self-critical style) compared with
ED patients without NSSI. In client-centred and experiential therapy, the critical self-talk or the inner self-criticising voice is known
as the inner critic (Stinckens, Lietaer & Leijssen, 2002). Therefore, from a phenomenological and psychotherapeutic viewpoint,
we may interpret self-injury in ED patients as the scars of the
inner critic. To prevent the patients from harming themselves,
either through eating-disordered or self-injurious behaviour, they
should become aware of the destructive power of their self-critical
attitude. Next, they should explore its dynamics, including the
internalisation of PPC. This awareness may be directly helpful
during the therapeutic process itself (Soenens, Luyckx, et al.,
2008; Soenens, Vansteenkiste, et al., 2008). ED patients should
learn to replace their self-harm by more healthy behaviours while
dealing with their (self-critical) emotional experiences. Finally,
they should improve their self-esteem and develop a positive
self-image as expressed in the care and preservation of a healthy
body.
Conclusion
This study shows that ED patients with ECP are more likely to
hurt themselves, thereby leaving scares as physical signals of
self-injury on their bodies. Presumably, the inner critic that is
the characteristic of those patients leads them to engage in NSSI
as a way to punish and torture themselves for their failures and/
or as a way to cope with their frustrating experiences. Finally,
our ndings suggest that PPC represents a risk factor for NSSI
in ED patients through the development of a self-critical perfectionist orientation.
Author.
50, 10031020.
Bulik, C. M., Tozzi, F., Anderson, C., Mazzeo, S. E., Aggen, S., & Sullivan, P. F. (2003). The relation between eating disorders and
Eur. Eat. Disorders Rev. (2011) 2011 John Wiley & Sons, Ltd and Eating Disorders Association.
L. Claes et al.
Soenens, B., Luyckx, K., Vansteenkiste, M., Luyten, P., Duriez, B., &
47, 106115.
Favaro, A., & Santonastaso, P. (1998). Impulsive and compulsive
186, 157165.
Favaro, A., & Santonastaso, P. (2000). Self-injurious behavior in anorexia nervosa. Journal of Nervous and Mental Disease, 188,
537542.
Flett, G. L., Hewitt, P. L., Oliver, J. M., & Macdonald, S. (2002). Perfectionism in children and their parents: a developmental analy-
Stinckens, N., Lietaer, G., & Leijssen, M. (2002). The inner critic on
10451056.
Klonsky, E. D. (2007). The functions of deliberate self-injury: a review of the evidence. Clinical Psychology Review, 27, 226239.
McCranie, E. W., & Bass, J. D. (1984). Childhood family antecedents
of dependency and self-criticism: implications for depression.
Journal of Abnormal Psychology, 93, 38.
Muehlenkamp, J. J., Engel, S. G., Wadeson, A., Crosby, R. D.,
48, 5259.
Eur. Eat. Disorders Rev. (2011) 2011 John Wiley & Sons, Ltd and Eating Disorders Association.