FULL-LENGTH REPORT Journal of Behavioral Addictions 5(1), pp.

77–89 (2016)
DOI: 10.1556/2006.5.2016.004

Maladaptive perfectionism as mediator among psychological control, eating

disorders, and exercise dependence symptoms in habitual exerciser
SEBASTIANO COSTA1*, HEATHER A. HAUSENBLAS2, PATRIZIA OLIVA1,
FRANCESCA CUZZOCREA1 and ROSALBA LARCAN1
1
Department of Human and Social Sciences, University of Messina, Messina, Italy
2
Department of Kinesiology, Jacksonville University, Jacksonville, FL, USA

(Received: June 17, 2015; revised manuscript received: October 28, 2015; accepted: December 26, 2015)

Background and aims: The current study examined the mediating role of maladaptive perfectionism among parental
psychological control, eating disorder symptoms, and exercise dependence symptoms by gender in habitual
exercisers. Methods: Participants were 348 Italian exercisers (n = 178 men and n = 170 women; M age =
20.57, SD = 1.13) who completed self-report questionnaires assessing their parental psychological control,
maladaptive perfectionism, eating disorder symptoms, and exercise dependence symptoms. Results: Results of the
present study confirmed the mediating role of maladaptive perfectionism for eating disorder and exercise dependence
symptoms for the male and female exercisers in the maternal data. In the paternal data, maladaptive perfectionism
mediated the relationships between paternal psychological control and eating disorder and exercise dependence
symptoms as full mediator for female participants and as partial mediator for male participants. Discussion: Findings
of the present study suggest that it may be beneficial to consider dimensions of maladaptive perfectionism and
parental psychological control when studying eating disorder and exercise dependence symptoms in habitual
exerciser.

Keywords: psychological control, maladaptive perfectionism, eating disorder, exercise dependence, habitual
exerciser

Exercise dependence is a term used to quantify and describe manifested by the reduction of social, occupational, or
pathological behaviors and attitudes related to exercise that recreational activities because of physical activity
are characterized by a preoccupation with exercise routines, (e.g., exercising rather than spending time with family and
withdrawal symptoms when unable to exercise, and an friends). Continuity is a persistence of exercise despite
interference with social relationships and occupational com- recurring negative physical or psychological effects
mitments (Berczik et al., 2012; Schreiber & Hausenblas, (e.g., continued exercising despite an overuse injury).
2015). In short, exercise dependence is a maladaptive pattern There is a strong link between exercise dependence and
of moderate to vigorous excessive exercise that manifests in eating disorders, with eating disorders being the most
negative physiological, psychological, and cognitive symp- common disorder to co-occur with exercise dependence
toms (Hausenblas & Symons Downs, 2002a). (Freimuth, Moniz, & Kim, 2011). Freimuth et al. (2011)
Based on the Diagnostic and statistical manual of mental noted that about 39–48% of people suffering from an eating
disorders-IV-R (DSM-IV-R) criteria for substance disorder also are exercise dependent (Bamber, Cockerill,
dependence (American Psychiatric Association, 2000), Rodgers, & Carroll, 2003; Klein et al., 2004). Because of the
Hausenblas and Symons Downs (2002a) identified the negative physical, psychological, and social outcomes and
following seven criteria or symptoms, at least three of which high co-occurrence of both exercise dependence and eating
must be met for the diagnosis of exercise dependence. First, disorders it is important to examine variables that predict
tolerance is the need for increasing the duration, frequency, symptoms of both of these disorders (Adams, 2009; Costa &
and intensity of exercise to obtain the desired effect or by Oliva, 2012; Gardner, Stark, Friedman, & Jackson, 2000).
experiencing less effect with the extended use of the same Knowledge of these specific characteristics will aid
amount of the exercise. Withdrawal is characterized by in the development and implementation of prevention and
experiencing unpleasant symptoms (such as anxiety, treatment interventions.
depression, and fatigue) when exercise stops or by using As well, given the strong relationship between exercise
exercise as a way to relieve or prevent these withdrawal dependence and eating disorders, certain etiological factors
symptoms. Intention effects occur when exercise behavior is may be common among these pathologies. In particular,
greater than what was originally intended. Loss of control is
the inability to reduce exercise, despite the desire to do so.
Time refers to a relatively large amount of time spent in * Corresponding author: Sebastiano Costa; Section of Psychology,
exercise (e.g., vacations are exercise-related) or exercise- Department of Human and Social Sciences, University of Messina,
related activities. Reductions in other activities are Via Bivona, 98122, Messina, Italy; E-mail: scosta@unime.it

ISSN 2062-5871 © 2016 Akadémiai Kiadó, Budapest

 Costa et al.

researchers have demonstrated the role of perfectionism and Perfectionism, in fact, is a multidimensional construct
parental control in the etiology and course of eating dis- comprising both maladaptive and adaptive features
orders (Polivy & Herman, 2002; Soenens et al., 2008). (Dunkley, Blankstein, Masheb, & Grilo, 2006). A central
However, there is no indication as to whether the interaction feature of perfectionism is the setting of high achievement
between parental psychological control and maladaptive standards, which in itself, is not pathological. However,
perfectionism would typify those identified as being exer- previous studies have shown that perfectionists with high
cise dependent. personal standards tend to have strong negative self-evalu-
Parental psychological control is an intrusive parenting ation to their unsatisfactory performance, that could create a
tactics to control the child's emotional state and to make rigid adherence to unrealistic standards (Frost, Marten,
children act, fell and think in parentally approved ways. Such Lahart, & Rosenblate, 1990; Shafran & Mansell, 2001;
intrusive tactics include love withdrawal, guilt-induction, Soenens et al., 2008). Previous studies suggest a central
conditional approval, shaming, instilling anxiety, and thinking role for maladaptive perfectionism with the etiology and
patterns (Barber, 1996). Parental psychological control repre- course of eating disorders and exercise dependence (Hall
sents a risk factor for maladjustment during different life et al., 2009; Shafran & Mansell, 2001). In according with
period, including adolescence and emerging adults (Barber, Soenens et al. (2008), driven by their sense of incompetence
Stolz, & Olsen, 2005; Costa, Hausenblas, Oliva, Cuzzocrea, and lack of control, maladaptive perfectionists could devel-
& Larcan, 2015; Costa, Soenens, Gugliandolo, Cuzzocrea, op a rigid focus on eating and exercise, and would engage in
& Larcan, 2015; Gugliandolo, Costa, Cuzzocrea, & Larcan, an escalating pattern of disordered eating and exercise
2014; Soenens & Vansteenkiste, 2010; Soenens, Vansteenkiste, dependence behaviors (Flett, Hewitt, Blankstein, & Mosher,
Luyten, Duriez, & Goossens, 2005). 1995; Hall, Kerr, Kozub, & Finnie, 2007; Shafran, Cooper, &
Parental psychological control may have a relevant role Fairburn, 2002).
in the sport context and it may be related to unhealthy In summary, research examining perfectionism as an
behaviors (Barber & Harmon, 2002; Costa, Hausenblas, intervening variable in associations between perceived
et al., 2015; Miller, 2012; Soenens et al., 2008; Soenens, parental control and eating disorder and exercise depen-
Vansteenkiste, & Luyten, 2010). Perceived parental psy- dence symptoms is needed for several reasons. First,
chological control, in fact, is related to both eating disorders perfectionism is an important risk factor that possibly
and exercise dependence (Costa, Hausenblas, et al., 2015). interacts with other risk factors to predict eating disorder
That is, parental psychological control is characterized by and exercise dependence symptoms (Boone, Soenens,
contingent expression of love, based on the performance of Braet, & Goossens, 2010; Hall et al., 2009). Second, several
children, creating negative self-evaluations with the con- studies propose that controlling parenting may promote
sequence of inducing rigid eating behaviors and exces- perfectionist concerns (Enns, Cox, & Larsen, 2000; Flett,
sive exercise (Soenens & Vansteenkiste, 2010; Soenens, Madorsky, Hewitt, & Heisel, 2002; Soenens et al., 2008).
Vansteenkiste, et al., 2005). For this reason, maladaptive In according with Soenens et al. (2008) psychologically
perfectionism could have a relevant role in the relation controlling parents would emphasize the attitude of their
between parental psychological control and exercise depen- children to criticize themselves with feelings of guilt when
dence and eating disorders. Soenens et al. (2008) demon- failing to meet their standards, creating also the idea in their
strated that maladaptive perfectionism mediated the links children that failure is unacceptable and that love of parents
between psychological control and eating disorder symp- depends on the respect of the rules, risking the development
toms. Eating disorders share numerous aspects with exercise of a maladaptive perfectionism orientation (Flett et al., 2002;
dependence symptoms and it is plausible that they have a Frost et al., 1990). Fourth, a maladaptive perfectionist
similar etiological process. Although the model suggested orientation would, in turn, create a vulnerability to disor-
by Soenens et al. (2008) was not verified in an exercise dered eating behaviors (Soenens et al., 2008) and excessive
context, it is reasonable that perfectionism could have a role exercise and mediate the relations between psychologically
as an intervening variable also in relationship to parental controlling parenting and eating disorder and exercise de-
psychological control and exercise dependence symptoms. pendence. Finally, identification of appropriate risk factors,
Integrating efforts to understand common etiology of exer- integrating environmental (i.e., parental psychological
cise dependence and eating disorders could have the eco- control) and individual (i.e., maladaptive perfectionism)
nomic efficiency of addressing two conditions within a variables, for the condition being targeted is essential to
single prevention intervention and also increase our under- developing effective prevention interventions (Kraemer,
standing of the onset and maintenance of problematic Stice, Kazdin, Offord, & Kupfer, 2001).
behaviors in an exercise context. Furthermore, despite that Gender is a very important factor in both eating patholo-
exercise dependence is related to parental psychological gy and exercise dependence. Women are more likely to
control (Costa, Hausenblas, et al., 2015) and perfectionism develop eating disorders than men, while women are less
(Costa, Coppolino, & Oliva, 2015; Hall, Hill, Appleton, & likely to be exercise dependent than men (Costa, Hausenblas,
Kozub, 2009; Symons Downs, Hausenblas, & Nigg, 2004), Oliva, Cuzzocrea, & Larcan, 2013; Hoek, 2006). However,
the role of both etiological factors (parental psychological some relationships among psychological control, perfec-
control and perfectionism) has not been investigated togeth- tionism, eating disorder and exercise dependence are incon-
er with exercise dependence. Understanding the role of sistent in male and female samples (Costa, Hausenblas,
perfectionism, may help identify those most at-risk for et al., 2013; Miller & Mesagno, 2014; Paulson & Rutledge,
exercise dependence and how parental psychological con- 2014; Soenens et al., 2008). For this reason, the present
trol may exacerbates exercise dependence symptoms. study aims to explore the relationship among exercise

78 | Journal of Behavioral Addictions 5(1), pp. 77–89 (2016)

 Exercise dependence perfectionism parental control

dependence, eating disorder, maladaptive perfectionism, psychological control. Barber (1996) provided evidence for
and parental psychological control symptoms by gender. the validity and one-dimensional factor structure of this
Consistent with previous research, it was hypothesized that scale. Subjects responded on a 3-point Likert-type scale
higher parental psychological control will predict increased ranging from 1 “not like her (him)” to 3 “a lot like her (him)”
exercise dependence and eating disorder symptoms (Costa, with higher scores reflecting greater perception of parental
Hausenblas, et al., 2015; Soenens et al., 2008). It was also psychological control. Subjects rated psychological control
hypothesized that higher parental psychological control will for both parents separately to examine the individual con-
predict increased use of maladaptive perfectionism and the tribution of mothers and fathers. The reliability and validity
use of maladaptive perfectionism should then predict in- of PCS-YSR have been demonstrated in cross-cultural
creased exercise dependence and eating disorder symptoms. research (Barber et al., 2005) and this scale has adequate
That is, maladaptive perfectionism should mediate the direct psychometric properties (Barber, 1996; Barber et al., 2005).
relationship among parental psychological control, exercise The Italian version of this instrument is widely used in
dependence symptoms, and eating disorder symptoms. research (Filippello, Sorrenti, Buzzai, & Costa, 2015;
Finally, it was hypothesized that, despite mean-level differ- Gugliandolo et al., 2014).
ences on some of the study variables (e.g., women report Maladaptive perfectionism. The Concern over Mistakes
higher eating disorder symptoms than men), the relationship Scale (nine items, e.g., “People will probably think less of
between the study constructs would be consistent across me if I make a mistake”) and the Doubts about Actions Scale
gender. (four items, e.g., “Even when I do something very carefully,
I often feel that it is not quite right”) from the Italian version
of the Frost Multidimensional Perfectionism Scale (MPS;
METHODS Frost et al., 1990) were used to measure maladaptive
perfections (Bieling, Israeli, & Antony, 2004; Frost et al.,
Participants 1990). A maladaptive perfectionism scale was constructed
by computing the mean of the items tapping Concern over
Participants were 348 Italian habitual exercisers that had Mistakes and Doubts about Actions with a five-point
been exercising for more than 6 months (n = 178 men and response scale ranging from 1 (strongly disagree) to 5
n = 170 women; age range = 19 to 22 years of age; M = (strongly agree) (Soenens, Elliot, et al., 2005; Soenens,
20.57, SD = 1.13; body mass index: M = 22.63, SD = Vansteenkiste, et al., 2005; Soenens, Vansteenkiste, Duriez,
3.35). The participants were mainly recruited from gyms & Goossens, 2006). For the Italian version of the MPS
and sport teams and spent a mean of 7.30 hours per week (Lombardo, 2008), items on the subscales Concern over
training (SD = 4.81). Most of the participants (n = 323) Mistakes (CM) and Doubts about Actions (D) loaded into a
had a high school diploma, 16 had a lower secondary unique factor, that represent the scales Concern over Mis-
education diploma, 8 had a university degree, and 1 partici- takes and Doubts about Actions (CMD). This scale has good
pant did not report this information. All the participants psychometric characteristics in various countries, including
came from two-parent families and they still lived with their Italy (Lombardo, 2008; Soenens, Elliot, et al., 2005;
parents. Soenens, Vansteenkiste, et al., 2005).
Eating disorder symptoms. The Italian version of the
Procedure Eating Attitudes Test-26 (EAT-26; Dotti & Lazzari, 1998;
Garner, Olmsted, Bohr, & Garfinkel, 1982) was used. The
Our convenient sample was recruited by soliciting volun- EAT-26 is a 26-item, self-report questionnaire that measures
teers through friends and appeals to gyms and sport associa- the level of disordered eating attitudes, thoughts, feelings, and
tions in Messina and Reggio Calabria. After describing the behaviors, using a six point Likert scale ranging from 0 (never)
study’s purpose, the inclusion criteria was presented and to 5 (always). The questionnaire includes three subscales:
participants that met the inclusion criteria and agreed to Dieting, Bulimia and Food Preoccupation, and Oral Control.
participate then signed the informed consent and voluntarily Scores greater than 20 indicate abnormal eating behaviors
completed the questionnaires individually or in small groups (Garner, Olmsted, & Polivy, 1983). The psychometric prop-
under the supervision of an experimenter. Inclusion criteria erties of the EAT are good and was widely used in research
were that exerciser had trained for at least 6 consecutive (Doninger, Enders, & Burnett, 2005; Koenig & Wasserman,
months, lived at home with both parents, and were at least 1995; Zmijewski & Howard, 2003).
18 years of age. Of the 424 participants who met the Exercise dependence symptoms. The Italian version of
inclusion criteria, 375 agreed to participate in the study. the Exercise Dependence Scale−R (EDS-R; Costa,
After signing the informed consent, 27 participants did not Cuzzocrea, Hausenblas, Larcan, & Oliva, 2012; Symons
complete the questionnaires. The final sample consisted of Downs et al., 2004) was used to measure exercise depen-
348 participants, representing a response rate of 82%. The dence symptoms. The EDS-R consists of 21 items scored on
questionnaires took about 30 minutes to complete. a 6-point Likert scale, ranging from 1 (never) to 6 (always).
Higher scores indicate more exercise dependence symp-
Measures toms. The instrument has seven subscales (three items for
each) based on the DSM-IV-TR criteria for substance
Psychological control. The 8-item Italian version of the dependence and a total score. The subscales are: (a) With-
Psychological Control Scale–Youth Self-Report (PCS- drawal (e.g., “I exercise to avoid feeling anxious”); (b)
YSR; Barber, 1996) was used to measure perception of Continuance (e.g., “I exercise despite recurring physical

Journal of Behavioral Addictions 5(1), pp. 77–89 (2016) | 79

 Costa et al.

problems”); (c) Tolerance (e.g., “I continually increase my lower than .08 and values of RMSEA that are equal or lower
exercise intensity to achieve the desired effect/benefits”); (d) than .06 indicate sufficient model fit.
Lack of control (e.g., “I am unable to reduce how long I
exercise”); (e) Reduction in other activities (e.g., “I would Ethics
rather exercise than spend time with family/friends”); (f)
Time (e.g., “I spend most of my free time exercising”); and Ethical principles were carried out in accordance with the
(g) Intention effects (e.g., “I exercise longer than I intend”). Declaration of Helsinki. The Institutional Review Board of
The criteria established in the DSM-IV-TR was utilised to the University of Messina, Messina, Italy approved the
classify individuals in the following exercise dependence study. All participants were informed about the study and
groups: at-risk for exercise dependence (i.e., scores of 5–6 provided informed consent before participating in the study
on average on the Likert scale in at least three of the seven procedures.
criteria), non-dependent symptomatic (i.e., scores of 3–4 on
average in at least three criteria, or scores of 5–6 on average
combined with scores of 3–4 on average in three criteria, but RESULTS
failing to meet the criteria of at-risk conditions), and non-
dependent asymptomatic (i.e., scores of 1–2 on average in at Preliminary analysis
least three criteria). The psychometric properties of the of
the Italian EDS-R are excellent (Costa et al., 2012; Costa & Based on the EDS-R and EAT-26 criteria, 11 participants
Oliva, 2012; Oliva, Costa, & Larcan, 2013). (3%; 5 men and 6 women) were classified as at-risk for both
exercise dependence and eating disorder, 11 participants
Data analysis (3%; 8 men and 3 women) were classified as at-risk for
exercise dependence only, and 41 participants (12%; 14 men
Structural Equation Modelling (SEM) with latent variables and 27 women) were classified as at-risk only for eating
was utilized to examine whether maladaptive perfection- disorder (see Table 1 for descriptive information). Prelimi-
ism could account for (i.e., mediate) the relationship nary analyses revealed that the men scored significantly
between parental psychological control, eating disorder higher than the women on exercise dependence symptoms,
symptoms, and exercise dependence symptoms by gender. t(346) = 2.75, p < .01. The women scored significantly
The models included latent variables for maladaptive higher than the men on eating disorder attitudes, t(346) =
perfectionism, parental psychological control, eating dis- 3.23, p < .01. No significant gender differences were evi-
order symptoms, and exercise dependence symptoms. Both denced for maladaptive perfections, psychological control
paternal and maternal psychological control and maladap- mother, and psychological control father. The five con-
tive perfectionism comprised three parcels. Instead of structs were all positively related (see Table 2).
using separate items as indicators, three parcels (groups)
of items was aggregate and used as indicators of the latent Mediation analysis for the male participants
constructs. Several studies have shown that parcelling has
some advantages relative to the use of individual items. First, a predictor model was estimated to test the direct paths
Individual parcels are likely to have a stronger relation to from the predictor (i.e., parental psychological control) to
the latent factor and are less likely to be influenced the outcome variable (i.e., eating disorder and exercise
by method effects, furthermore parcelling results in a dependence) in the male sample, including correlations
smaller number of indicators per latent factor and is more between the error terms for exercise dependence and eating
likely to meet the assumptions of normality (Coffman & disorder symptoms. This model did not include maladaptive
MacCallum, 2005; Little, Cunningham, Shahar, & Widaman, perfectionism, to verify the first step of the mediation
2002; Marsh, Hau, Balla, & Grayson, 1998). Analysis of the process, in which an acceptable fit of a model comprising
covariance matrices was conducted using EQS 6.2 and only direct paths between the predictor variables and the
solutions were generated on the basis of maximum- outcome is confirmed. Estimation of the maternal model,
likelihood estimation. χ2(62) = 127.60, p < .001; CFI = .93; NNFI = .92;
To explore the mediating role of need satisfaction, the SRMR = .06; RMSEA = .08 (90% CI = .06–.10), showed
SEM approach advanced by Baron and Kenny (1986), a significant path from psychological control to eating
Holmbeck (1997), and Shrout and Bolger (2002) for testing disorder (β = .23; p < .01), and exercise dependence
mediation was used. Additionally, bootstrapping was used (β = .25; p < .01). Similarly, estimation of the paternal
to estimate the standard errors (SEs) and 95% bias-corrected model, χ2(62) = 124.76, p < .01; CFI = .93; NNFI = .92;
confidence intervals (CIs) for all model estimates (Shrout & SRMR = .06, RMSEA = .08 (90% CI = .06–.10), showed
Bolger, 2002). Hu and Bentler (1999) argued for using a significant path from psychological control to eating
combinations of cut-off values to examine model fit. Ac- disorder (β = .36; p < .01), and exercise dependence
cordingly, we examined the chi-square statistic, and fit (β = .37; p < .01).
indices such as the Comparative Fit Index (CFI), Non- Second, a full mediation model with psychological con-
Normed Fit Index (NNFI), the Standard Root Mean Square trol was tested, that is, a model in which psychological
Residual (SRMR) and the Root Mean Square Error of control was related only indirectly to eating disorder and
Approximation (RMSEA). Hu and Bentler (1999) proposed exercise dependence through maladaptive perfectionism,
that values of CFI that are equal or greater than .90, values of including correlations between the error terms for exercise
NNFI that are equal or greater than .90, values of SRMR dependence and eating disorder symptoms. Estimation of

80 | Journal of Behavioral Addictions 5(1), pp. 77–89 (2016)

 Exercise dependence perfectionism parental control

Table 1. Mean (M), standard deviation (SD), and Cronbach’s alpha (α) for the male and female samples
Male Female

α M SD α M SD
1. Psychological control mother .83 1.61 .47 .82 1.62 .47
2. Psychological control father .80 1.52 .45 .79 1.58 .46
3. Maladaptive perfectionism .93 2.59 .76 .88 2.58 .66
4. Eating disorder .92 2.43 1.02 .89 2.73** .97
5. Exercise dependence .95 11.66** 13.02 .94 7.67 9.89
Note: ** value significantly higher between gender for p < .01.

Table 2. Pearson’s correlational coefficients of the variables

1 2 3 4 5
1. Psychological control mother − .53** .59** .27** .29**
2. Psychological control father .45** − .41** .24** .17*
3. Maladaptive perfectionism .46** .37** − .43** .42**
4. Eating disorder .20** .31** .36** − .52**
5. Exercise dependence .20** .29** .34** .37** −
Note: *p < .05; **p < .01. Lower diagonal correlation matrix of the male data; upper diagonal correlation matrix of the female data.

this model (see Figure 1) yielded acceptable fit for both the (β = .11; p > .05), and exercise dependence (β = .10;
maternal data, χ2(100) = 168.49, p < .01; CFI = .94; p > .05) was no longer significant after including maladap-
NNFI = .93; SRMR = .06, RMSEA = .06 (90% tive perfectionism as a mediator. Instead in the paternal
CI = .05–.08), and the paternal data, χ2(100) = 176.89, model, after including maladaptive perfectionism as a
p < .01; CFI = .93; NNFI = .92; SRMR = .09, RMSEA = mediator, the paths from psychological control to eating
.07 (90% CI = .05–.08). Psychological control was related disorders (β = .28; p < .01) and to exercise dependence
positively to maladaptive perfectionism in both the maternal remained significant (β = .28; p < .05).
(β = .49; p < .01) and paternal model (β = .42; p < .01). In In summary, the full mediation model was the most
turn, maladaptive perfectionism was related positively to parsimonious and best fitting model for the maternal data
eating disorders (β = .34; p < .01), and exercise dependence (see Figure 1), whilst the partially mediated model was the
(β = .33; p < .01) in the maternal model. Maladaptive per- best fitting model for the paternal data (see Figure 2). In this
fectionism was also related positively to eating disorders model, the direct paths from psychological control to eating
(β = .35; p < .01), and exercise dependence (β = .35; disorders and exercise dependence were not significant after
p < .01) in the paternal model. including the mediator in the maternal model, while they
Third, a partial mediation model was assessed by were significant after including the mediator in the paternal
adding a direct path from psychological control to model. Furthermore, the indirect relation of psychological
eating disorder and exercise dependence whilst controlling control with eating disorders and exercise dependence
for maladaptive perfectionism, including correlations mediated by maladaptive perfectionism were statistically
between the error terms for exercise dependence and eating significant for both maternal and paternal ratings, suggesting
disorder symptoms. Estimation of this model yielded that perfectionism was a full mediator in the maternal model
acceptable fit both for the maternal data, χ2(98) = while it represented a partial mediation in the paternal model
167.01, p < .01; CFI = .94; NNFI = .93; SRMR = .06, (see Table 2).
RMSEA = .06 (90% CI = .05–.08), and for the paternal
data χ2(98) = 165.48, p < .01; CFI = .94; NNFI = .93; Mediation analysis for the female participants
SRMR = .06, RMSEA = .06 (90% CI = .05–.08).
This model did not provide a significantly better fit than First, a model estimating the direct paths from the predictor
the full mediation model in the maternal ratings [Δχ2(2) = (i.e., parental psychological control) to the outcome variable
1.38; p > .05]. In the paternal ratings instead the partial (i.e., eating disorder and exercise dependence) in the female
mediation model does not provide a significantly better fit sample was tested, including correlations between the
than the full mediation model [Δχ2(2) = 9.41, p < .05], error terms for exercise dependence and eating disorder
suggesting that the full mediation model provided the most symptoms. This model did not include maladaptive perfec-
parsimonious representation of the data only for the mater- tionism. Estimation of the maternal model, χ2(62) =
nal model. 168.53, p < .001; CFI = .91; NNFI = .88; SRMR = .07,
Moreover, in the maternal model, the originally signifi- RMSEA = .10 (90% CI = .08–.12), showed a significant
cant path from psychological control to eating disorder path from psychological control to eating disorder (β = .37;

Journal of Behavioral Addictions 5(1), pp. 77–89 (2016) | 81

 Costa et al.

Figure 1. Full mediation models between psychological control, maladaptive perfectionism, eating disorder and exercise dependence.
Note: Coefficients shown are standardized path coefficients. The first coefficient shown is for the maternal psychological control
model in a male sample. The second coefficient shown is for the maternal psychological control model in a female sample. The third
coefficient shown is for the paternal psychological control model in a female sample

p < .01), and exercise dependence (β = .26; p < .01). Third, a partial mediation model by adding a direct path
Similarly, estimation of the paternal model, χ2(62) = from psychological control to eating disorder and exercise
165.55, p < .01; CFI = .91; NNFI = .88; SRMR = .07, dependence whilst controlling for maladaptive perfection-
RMSEA = .09 (90% CI = .08–.12), showed a significant ism and that included correlations between the error terms
path from psychological control to eating disorder (β = .23; for exercise dependence and eating disorder symptoms was
p < .01), and exercise dependence (β = .28; p < .01). assessed. Estimation of this model yielded acceptable fit
Second, a full mediation model with psychological con- both for the maternal data, χ2(98) = 204.90, p < .01;
trol was tested, that is, a model in which psychological CFI = .93; NNFI = .91; SRMR = .07, RMSEA = .08
control was related only indirectly to eating disorders and (90% CI = .07–.10), and for the paternal data χ2(98) =
exercise dependence through maladaptive perfectionism, 213.94, p < .01; CFI = .92; NNFI = .90; SRMR = .07,
including correlations between the error terms for exercise RMSEA = .08 (90% CI = .07–.10).
dependence and eating disorder symptoms. Estimation This model did not provide a significantly better fit than
of this model (see Figure 1) yielded acceptable fit for both the full mediation model in the maternal [Δχ2(2) = 1.65;
the maternal data, χ2(100) = 206.55, p < .01; CFI = .93; p > .05], and the paternal ratings [Δχ2(2) = 0.89;
NNFI = .91; SRMR = .07, RMSEA = .08 (90% p > .05], suggesting that the full mediation model provided
CI = .06–.09), and the paternal data, χ2(100) = 214.83, the most parsimonious representation of the data.
p < .01; CFI = .92; NNFI = .90; SRMR = .07, Moreover, in the maternal model, the originally signifi-
RMSEA = .08 (90% CI = .07–.10). Psychological control cant path from psychological control to eating disorder
was related positively to maladaptive perfectionism in both (β = .07; p > .05), and exercise dependence (β = .09;
the maternal (β = .63; p < .01) and the paternal model p > .05) was no longer significant after including maladap-
(β = .40; p < .01). In turn, maladaptive perfectionism was tive perfectionism as a mediator. Similarly to the maternal
related positively to eating disorders (β = .51; p < .01), model, in the paternal model, the originally significant paths
and exercise dependence (β = .49; p < .01) in the from psychological control to eating disorder (β = .04;
maternal model. Maladaptive perfectionism was also related p > .05) and exercise dependence (β = .09; p > .05) were
positively to eating disorders (β = .51; p < .01), and exer- no longer significant after including maladaptive perfection-
cise dependence (β = .50; p < .01) in the paternal model. ism as a mediator.

82 | Journal of Behavioral Addictions 5(1), pp. 77–89 (2016)

 Exercise dependence perfectionism parental control

Figure 2. Partial mediation model between paternal psychological control, maladaptive perfectionism, eating disorder and exercise
dependence in a male sample.
Note: Coefficients shown are standardized path coefficients for the paternal psychological control model in a male sample

In summary, the full mediation model was the most R-CFI = .93; R-RMSEA = .06 (90% CI = .05–.07).
parsimonious and best fitting model for the maternal data Similar results were provided for the paternal model, where
and the paternal data (see Figure 1). In this model, the direct the fit indices of the constrained model, χ²(214) = 413.58,
paths from psychological control to eating disorders and S-B χ²(214) = 354.45, p < .05, R-CFI = .93; R-RMSEA =
exercise dependence were not significant after including the .06 (90% CI = .05–.07), did not significantly differ from the
mediator in both the maternal and paternal models, suggest- unconstrained model, χ²(196) = 379.41, S-B χ²(196) =
ing that perfectionism was a full mediator in both of the 329.16, p < .05, R-CFI = .93; R-RMSEA = .06 (90%
models. While the indirect relation of psychological control CI = .05–.07), indicating measurement equivalence across
with eating disorders and exercise dependence mediated by gender both for the maternal [Δχ2(18) = 28.46, p > .05;
maladaptive perfectionism were statistically significant for ΔCFI = .006] and paternal model [Δχ2(18) = 25.87,
both the maternal and paternal ratings (see Table 3 and p > .05; ΔCFI = .005].
Table 4).

Multiple group analysis

DISCUSSION

As a subsidiary to the main analyses, we tested the equality Researchers have highlighted the necessity of exploring
of the model parameters across gender. Specifically, we the mechanisms that establish the relationship between
compared the revised model unconstrained across gender psychological control and personal functioning (Barber &
with a nested model in which all factor loadings, path Harmon, 2002; Costa, Soenens, et al., 2015; Gugliandolo
coefficients, and the error covariances of the two partial et al., 2014; Soenens, Vansteenkiste, et al., 2005). Costa,
mediation models were constrained to be invariant across Hausenblas, et al. (2015) found that parental psychological
gender. The fit indices of the constrained maternal model, control may have an effect on the development of exercise
χ2(214) = 408.22; S-B χ²(214) = 355.97, p < .05, dependence in habitual exercisers. Several studies have
R-CFI = .93; R-RMSEA = .06 (90% CI = .05–.07), did shown that psychological control create a vulnerability to
not significantly differ from the unconstrained model, maladaptive patterns of development through a perfection-
χ2(196) = 371.90; S-B χ²(196) = 327.70, p < .05, ism orientation (Soenens et al., 2006; Soenens et al., 2008).

Journal of Behavioral Addictions 5(1), pp. 77–89 (2016) | 83

 Costa et al.

Table 3. Path estimates, SEs and 95% CIs for models in the male sample
β B-SE Lower bound (BC) 95% CI Upper bound (BC) 95% CI
Maternal model
Direct effect
Psychological Control → Maladaptive Perfectionism .48 .08 .30 .62
Psychological Control → Eating Disorder .11 .13 −.14 .34
Psychological Control → Exercise Dependence .11 .11 −.14 .33
Maladaptive Perfectionism → Eating Disorder .28 .13 .03 .51
Maladaptive Perfectionism → Exercise Dependence .28 .11 .07 .49
Indirect effect via maladaptive perfectionism
Psychological Control → Eating Disorder .13 .07 .02 .27
Psychological Control → Exercise Dependence .13 .06 .04 .28
Paternal model
Direct effect
Psychological Control → Maladaptive Perfectionism .40 .11 .19 .61
Psychological Control → Eating Disorder .29 .12 .05 .54
Psychological Control → Exercise Dependence .29 .11 .04 .48
Maladaptive Perfectionism → Eating Disorder .21 .10 .02 .43
Maladaptive Perfectionism → Exercise Dependence .21 .10 .01 .40
Indirect effect via maladaptive perfectionism
Psychological Control → Eating Disorder .09 .05 .02 .22
Psychological Control → Exercise Dependence .08 .05 .01 .22
Note: B-SE = bootstrapped standards errors; BC 95% CI = bias corrected-confidence interval.

Although research has yielded evidence for the intervening The first results of the present study indicate that there is
role of perfectionism in associations between parental control a significant relationship between parental psychological
and eating disorders (Soenens et al., 2008), this hypothesis has control and both exercise dependence and eating disorder
not been fully addressed in research on habitual exercisers and symptoms. These findings are similar to the findings of
on exercise dependence symptoms. The aim of our study, Soenens et al. (2008) and Costa, Hausenblas, et al. (2015),
therefore, was to examine both perceived parental psychologi- in that paternal psychological control was positively related
cal control and perfectionism in relation to eating disorder and to eating disorder and exercise dependence symptoms. The
exercise dependence symptoms. results of our study give further insight into these findings

Table 4. Path estimates, SEs and 95% CIs for models in the female sample
β B-SE Lower bound (BC) 95% CI Upper bound (BC) 95% CI
Maternal model
Direct effect
Psychological Control → Maladaptive Perfectionism .63 .06 .50 .75
Psychological Control → Eating Disorder .09 .16 −.21 .42
Psychological Control → Exercise Dependence −.07 .16 −.41 .24
Maladaptive Perfectionism → Eating Disorder .45 .14 .15 .70
Maladaptive Perfectionism → Exercise Dependence .54 .13 .27 .79
Indirect effect via maladaptive perfectionism
Psychological Control → Eating Disorder .28 .09 .10 .47
Psychological Control → Exercise Dependence .34 .10 .15 .57
Paternal model
Direct effect
Psychological Control → Maladaptive Perfectionism .40 .09 .22 .56
Psychological Control → Eating Disorder .07 .12 −.16 .32
Psychological Control → Exercise Dependence .10 .13 −.15 .38
Maladaptive Perfectionism → Eating Disorder .48 .10 .24 .66
Maladaptive Perfectionism → Exercise Dependence .45 .10 .24 .62
Indirect effect via maladaptive perfectionism
Psychological Control → Eating Disorder .19 .06 .10 .32
Psychological Control → Exercise Dependence .18 .06 .09 .31
Note: B-SE = bootstrapped standards errors; BC 95% CI = bias corrected-confidence interval.

84 | Journal of Behavioral Addictions 5(1), pp. 77–89 (2016)

 Exercise dependence perfectionism parental control

suggesting that both paternal and maternal psychological develo a "drive for thinness" (Hausenblas & Fallon, 2002;
control is related to eating disorder and exercise dependence McCreary & Sasse, 2000; Vocks, Hechler, Rohrig, &
symptoms. It is plausible that maternal psychological Legenbauer, 2009). Furthermore, the differing nature of
control could have a greater impact on eating disorder and each relationship between mother, father, sons and daugh-
exercise dependence symptoms indirectly, rather than ters might influence the link between psychological control
directly, through its relationship with other variables and adjustment (Soenens & Vansteenkiste, 2010; Steinberg,
(e.g., maladaptive perfectionism; Costa, Hausenblas, 2001). In our study separate analyses were conducted for
et al., 2015; Pole, Waller, Stewart, & Parkin‐Feigenbaum, male and female habitual exercisers and the effects of
1988; Soenens et al., 2008; Vandereycken, 1994). Second, psychological control were differentiated between maternal
and consistent with previous research, it was found that and paternal perceived psychological control. However
psychological control, both maternal and paternal, predicted structural associations were relatively analogous in maternal
maladaptive perfectionism (Soenens, Elliot, et al., 2005; and paternal models. This result was consistent with previ-
Soenens, Vansteenkiste, et al., 2005; Soenens et al., 2008), ous studies that examinated the dynamics of parental
while maladaptive perfectionism, predicted eating psychological control in fathers and mothers (Barber &
disorder symptoms (Bulik et al., 2003; Flett et al., 1995; Harmon, 2002). In fact, both for men and women the
Halmi et al., 2000; Soenens et al., 2008) and exercise relations between maternal psychological control and eating
dependence symptoms (Hagan & Hausenblas, 2003; disorder and exercise dependence symptoms were mediated
Hausenblas & Symons Downs, 2002b; Symons Downs by maladaptive perfectionism. The same results were found
et al., 2004). for maternal psychological control in male habitual exerci-
The current study also hypothesized an effect of mal- sers, while paternal psychological control in male habitual
adaptive perfectionism as a mediator in the relationship exercisers was partially mediated by maladaptive perfec-
between psychological control and eating disorder and tionism in its relation with eating disorder and exercise
exercise dependence symptoms. Results confirmed the me- dependence symptoms. This pattern suggests independent
diating role of maladaptive perfectionism as a full mediator additive effects. Consequently, habitual exercisers’ eating
for eating disorder and exercise dependence symptoms for disorder and exercise dependence symptoms were directly
both male and female habitual exercisers in the maternal predicted by paternal psychological control and indirectly
data. Furthermore, in the paternal data, maladaptive through the mediation of maladaptive perfectionism. Pater-
perfectionism mediated the relationships between paternal nal psychological control could promote maladaptive per-
psychological control and eating disorder and exercise fectionism leading to the development of eating disorder and
dependence symptoms as full mediator for the female exercise dependence symptoms. However, on the other
participants and as partial mediator for the male participants. hand, the association between parental psychological con-
In according with the suggestions of Soenens et al. (2008), trol and eating disorder and exercise dependence symptoms
maladaptive perfectionism, in fact, would be the result of a could also depend on not clear aspects that may be investi-
family context in which parents approve their child’s be- gated in future studies. In our view, the different pattern of
haviour only when meet their harsh standards and criticize relationships between the paternal and maternal models in
their child when fails. As a consequence children gradually men is mainly due to the fact that paternal psychological
engage in negative self-evaluations and learn to impose control seems to have a more pronounced and systematic
these high and rigid standards on themselves (Flett et al., association with eating disorder and exercise dependence
2002; Soenens et al. 2008). For this reason, the experience symptoms (Costa, Hausenblas, et al., 2015; Soenens et al.,
of psychological control would yield exercisers sensitive for 2008). In fact, even though fathers may be less involved in
the development of an enduring perfectionist orientation. In parenting than mothers overall, they may step in when more
habitual exercisers, perfectionism could be characterized by serious child-rearing problems arise and when conflict and
restrictiveness and desire for control. These typical over- negativity are more likely (May, Kim, McHale, & Crouter,
controlled personality features of perfectionism, in associa- 2006).
tion with the worries about being unable to adhere their The implications of these findings could be a very
standards for performances and body appearance, could lead important in prevention and treatment efforts. The deleteri-
to a rigid focus both on eating and exercise, with an ous effects of parental psychological control, both on eating
escalating pattern of disordered eating and excessive exer- disorder and on exercise dependence symptoms (Soenens &
cise behaviors (Hall, 2006; Shafran et al., 2002; Soenens Vansteenkiste, 2010), require the implementation of targeted
et al., 2008). interventions to educate parents in using fewer intrusions and
Finally, consistent with previous research, men reported parental autonomy support, as opposed to psychological
more exercise dependence symptoms than women, and control. It is important to educate parents about psychologi-
women reported more eating disorder symptoms than men cal control being a universally negative parenting strategy
(Berczik et al., 2012; Costa, Hausenblas, et al., 2013; and to help parents identify and reduce the use of such
Lewinsohn, Seeley, Moerk, & Striegel‐Moore, 2002; dysfunctional practices. Parent training could be used to
Striegel‐Moore et al., 2009). A possible explanation could help parents to get to practice autonomy supportive skills by
be that, from a body image perspective, exercise allow to taking part in various exercises from common situations in
meet in men the societal standard of muscular psysique, with familial daily living. Finally, results of this study provide
the risk to develop a "drive for muscularity". In female, evidence for factors of relevance for both exercise depen-
instead, exercise to yield their societal standard of psysique dence and eating disorders that could serve as focal points
should be accompanied by calorie reduction, with the risk to for integrated prevention interventions. Results of the study

Journal of Behavioral Addictions 5(1), pp. 77–89 (2016) | 85

 Costa et al.

suggest that the therapist and client could work together to RL assisted with data interpretation, manuscript editing, and
identify and correct the client’s maladaptive perfectionism study supervision. All authors take responsibility for the
thought processes that result in maladaptive eating or integrity of the data and the accuracy of the data analysis.
exercise behavior. This can involve helping a client to All authors contributed to and have approved the final
understand the development of his or her maladaptive manuscript.
perfectionism thought and how it can result in eating
disorder or exercise dependence symptoms. Conflict of interest: All authors report that they have no
The present study shows a number of shortcomings that conflicts of interest.
need to be addressed in future research. First, the data were
collected at one time point and not across multiple time
points as it should be done for a true mediational analysis.
Because of the cross-sectional and self-report design of our REFERENCES
study, we cannot establish either the direction or causality of
the effects. As well, the self-report data may have intrinsic Adams, J. (2009). Understanding exercise dependence. Journal of
limitations such as social desirability, so the use of objective Contemporary Psychotherapy, 39, 231–240. doi: 10.1007/
and multi-informant measures and experimental designs s10879-009-9117-5
are encouraged (Gugliandolo, Costa, Cuzzocrea, Larcan, American Psychiatric Association. (2000). Diagnostic and statis-
& Petrides, 2015). Although our sample consisted of habit- tical manual of mental disorders (4th ed.). Washington, DC:
ual exercisers, this sample was deemed justified, as eating Author.
disorder and exercise dependence symptoms are prevalent in Bamber, D. J., Cockerill, I. M., Rodgers, S., & Carroll, D. (2003).
this population (e.g., Costa, Hausenblas, et al., 2013) and Diagnostic criteria for exercise dependence in women. British
because psychological control is a risk factor for eating Journal of Sports Medicine, 37, 393–400. doi: 10.1136/
disorder and exercise dependence symptoms in this popu- bjsm.37.5.393
lation (e.g., Costa, Hausenblas, et al., 2015; Soenens et al. Barber, B. K. (1996). Parental psychological control: Revisiting a
2008). Future researchers are encouraged to rely on larger neglected construct. Child development, 67, 3296–3319.
and more heterogeneous samples, including clinical popula- Barber, B. K., & Buehler, C. (1996). Family cohesion and
tions, to explore the generalizability of the hypothesized enmeshment: Different constructs, different effects. Journal
model. of Marriage and Family, 58, 433–441.
Notwithstanding these limitations, our study findings Barber, B. K., & Harmon, E. L. (2002). Violating the self: Parental
offer an original contribution for examining the parents’ psychological control of children and adolescents. In B. K.
role in the development of eating disorder and exercise Barber (Ed.), Intrusive parenting: How psychological control
dependence symptoms in habitual exercisers. This is the first affects children and adolescents. Washington, DC: American
study that examines the mediational role of maladaptive Psychological Association.
perfectionism in the relation between parental psychological Barber, B. K., Stolz, H. E., & Olsen, J. A. (2005). Parental support,
control and eating disorder and exercise dependence symp- psychological control, and behavioral control: Assessing rele-
toms in a sample of habitual exercisers, showing that the vance across time, method, and culture. Monographs of the
intrusive and controlling parenting seems to generate a Society for Research in Child Development, 70, 1–137.
vulnerability to perfectionism and maladjustment psycho- Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator
pathology that in both male and female habitual exercisers variable distinction in social psychological research: Concep-
could lead in eating disorder and exercise dependence. tual, strategic, and statistical considerations. Journal of Per-
Moreover, our study furthers the examination of the role sonality and Social Psychology, 51, 1173–1182. doi: 10.1037/
of parental psychological control in exercise context. A 0022-3514.51.6.1173
more detailed understanding of the various mechanisms, Berczik, K., Szabó, A., Griffiths, M. D., Kurimay, T., Kun, B., Urbán,
individual and/or interactional, that are operating will pro- R., & Demetrovics, Z. (2012). Exercise addiction: Symptoms,
vide psychologists with a well-founded theoretical basis for diagnosis, epidemiology, and etiology. Substance Use & Misuse,
therapeutic intervention implementation, and support to 47, 403–417. doi: 10.3109/10826084.2011.639120
reduce the potential risk of experiencing unhealthy eating Bieling, P. J., Israeli, A. L., & Antony, M. M. (2004). Is perfec-
and exercise behaviors. tionism good, bad, or both? Examining models of the perfec-
tionism construct. Personality and Individual Differences, 36,
1373–1385. doi: 10.1016/S0191-8869(03)00235-6
Funding sources: No financial support was received for Boone, L., Soenens, B., Braet, C., & Goossens, L. (2010). An
this study. empirical typology of perfectionism in early-to-mid adoles-
cents and its relation with eating disorder symptoms. Behav-
Authors’ contribution: SC assisted with generation of the iour Research and Therapy, 48, 686–691. doi: 10.1016/j.
initial draft of the manuscript, data analyses, and manuscript brat.2010.03.022
editing; HH assisted with manuscript preparation, manu- Bulik, C. M., Devlin, B., Bacanu, S. A., Thornton, L., Klump,
script editing, interpretation of data, study design and K. L., Fichter, M. M., & Kaye, W. H. (2003). Significant
concept; PO assisted with manuscript preparation, manu- linkage on chromosome 10p in families with bulimia nervosa.
script editing, study design and concept; FC assisted with The American Journal of Human Genetics, 72, 200–207. doi:
data analysis, data interpretation, and manuscript editing; 10.1086/345801

86 | Journal of Behavioral Addictions 5(1), pp. 77–89 (2016)

 Exercise dependence perfectionism parental control

Coffman, D. L., & MacCallum, R. C. (2005). Using parcels to Freimuth, M., Moniz, S., & Kim, S. R. (2011). Clarifying exercise
convert path analysis models into latent variable models. addiction: Differential diagnosis, co-occurring disorders, and
Multivariate Behavioral Research, 40, 235–259. phases of addiction. International Journal of Environmental
Costa, S., Coppolino, P., & Oliva, P. (2015). Exercise dependence Research and Public Health, 8(10), 4069–4081.
and maladaptive perfectionism: The mediating role of basic Frost, R. O., Marten, P., Lahart, C., & Rosenblate, R. (1990). The
psychological needs. International Journal of Mental Health dimensions of perfectionism. Cognitive Therapy and Research,
and Addiction. doi: 10.1007/s11469-015-9586-6 14, 449–468. doi: 10.1007/bf01172967
Costa, S., Cuzzocrea, F., Hausenblas, H. A., Larcan, R., & Oliva, Gardner, R. M., Stark, K., Friedman, B. N., & Jackson, N. A.
P. (2012). Psychometric examination and factorial validity of (2000). Predictors of eating disorder scores in children
the exercise dependence scale–revised in Italian exercisers. ages 6 through 14: A longitudinal study. Journal of Psychoso-
Journal of Behavioral Addictions, 1, 186–190. doi: 10.1556/ matic Research, 49, 199–205. doi: 10.1016/s0022-3999(00)
JBA.1.2012.009 00172-0
Costa, S., Hausenblas, H. A., Oliva, P., Cuzzocrea, F., & Larcan, Garner, D. M., Olmsted, M. P., Bohr, Y., & Garfinkel, P. E. (1982).
R. (2015). Perceived parental psychological control and The eating attitudes test: Psychometric features and clinical
exercise dependence symptoms in competitive athletes. Inter- correlates. Psychological Medicine, 12, 871–878. doi: http://
national Journal of Mental Health and Addiction, 13, 59–72. dx.doi.org/10.1017/S0033291700049163
doi: 10.1007/s11469-014-9512-3 Garner, D. M., Olmstead, M. P., & Polivy, J. (1983). Development
Costa, S., Hausenblas, H. A., Oliva, P., Cuzzocrea, F., & Larcan, and validation of a multidimensional eating disorder inventory for
R. (2013). The role of age, gender, mood states and exercise anorexia nervosa and bulimia. International Journal of Eating
frequency on exercise dependence. Journal of Behavioral Disorders, 2, 15–34. doi: 10.1016/s0022-3999(00)00172-0
Addictions, 2, 216–223. doi: 10.1556/JBA.2.2013.014 Gugliandolo, M. C., Costa, S., Cuzzocrea, F., & Larcan, R.
Costa, S., & Oliva, P. (2012). Examining relationship between (2014). Trait emotional intelligence as mediator between
personality characteristics and exercise dependence. Review of psychological control and behaviour problems. Journal of
Psychology, 19, 5–12. Child and Family Studies, 24, 2290–2300. doi: 10.1007/
Costa, S., Soenens, B., Gugliandolo, M. C., Cuzzocrea, F., & s10826-014-0032-3
Larcan, R. (2015). The mediating role of experiences of need Gugliandolo, M. C., Costa, S., Cuzzocrea, F., Larcan, R., &
satisfaction in associations between parental psychological Petrides, K. V. (2015). Trait emotional intelligence and behav-
control and internalizing problems: A study among Italian ioral problems among adolescents: A cross-informant design.
college students. Journal of Child and Family Studies, 24, Personality and Individual Differences, 74, 16–21. doi:
1106–1116. doi: 10.1007/s10826-014-9919-2 10.1016/j.paid.2014.09.032
Doninger, G. L., Enders, C. K., & Burnett, K. F. (2005). Hagan, A. L., & Hausenblas, H. A. (2003). The relationship
Validity evidence for Eating Attitudes Test scores in a between exercise dependence symptoms and perfectionism.
sample of female college athletes. Measurement in Physical American Journal of Health Studies, 18, 133–137.
Education and Exercise Science, 9, 35–49. doi: 10.1207/ Hall, H. K. (2006). Perfectionism: A hallmark quality of world
s15327841mpee0901_3 class performers, or a psychological impediment to athletic
Dotti, A., & Lazzari, R. (1998) Validation and reliability of the development. Essential Processes for Attaining Peak Perfor-
Italian EAT-26. Eating and Weight Disorders, 3, 188–194. doi: mance, 1, 178–211.
10.1007/bf03340009 Hall, H. K., Hill, A. P., Appleton, P. R., & Kozub, S. A. (2009). The
Dunkley, D. M., Blankstein, K. R., Masheb, R. M., & Grilo, C. M. mediating influence of unconditional self-acceptance and labile
(2006). Personal standards and evaluative concerns dimensions self-esteem on the relationship between multidimensional per-
of “clinical” perfectionism: A reply to Shafran et al. (2002, fectionism and exercise dependence. Psychology of Sport and
2003) and Hewitt et al. (2003). Behaviour Research and Exercise, 10, 35–44. doi: 10.1016/j.psychsport.2008.05.003
Therapy, 44, 63–84. doi: 10.1016/j.brat.2004.12.004 Hall, H. K., Kerr, A. W., Kozub, S. A., & Finnie, S. B. (2007).
Enns, M. W., Cox, B. J., & Larsen, D. K. (2000). Perceptions Motivational antecedents of obligatory exercise: The influence
of parental bonding and symptom severity in adults with of achievement goals and multidimensional perfectionism.
depression: Mediation by personality dimensions. Canadian Psychology of Sport and Exercise, 8, 297–316. doi:
Journal of Psychiatry. Revue Canadienne de Psychiatrie, 45, 10.1016/j.psychsport.2006.04.007
263–268. Halmi, K. A., Sunday, S. R., Strober, M., Kaplan, A., Woodside,
Filippello, P., Sorrenti, L., Buzzai, C., & Costa, S. (2015). Per- D. B., Fichter, M., & Kaye, W. H. (2000). Perfectionism in
ceived parental psychological control and learned helplessness: anorexia nervosa: Variation by clinical subtype, obsessionality,
The role of school self-efficacy. School Mental Health, 7, 298– and pathological eating behavior. American Journal of
310. doi: 10.1007/s12310-015-9151-2 Psychiatry, 157, 1799–1805. doi: 10.1176/appi.ajp.157.
Flett, G. L., Hewitt, P. L., Blankstein, K. R., & Mosher, S. W. 11.1799
(1995). Perfectionism, life events, and depressive symptoms: A Hausenblas, H. A., & Fallon, E. A. (2002). Relationship among
test of a diathesis-stress model. Current Psychology, 14, body image, exercise behavior, and exercise dependence symp-
112–137. doi: 10.1007/bf02686885 toms. International Journal of Eating Disorders, 32, 179–185.
Flett, G. L., Madorsky, D., Hewitt, P. L., & Heisel, M. J. (2002). doi: 10.1002/eat.10071
Perfectionism cognitions, rumination, and psychological dis- Hausenblas, H. A., & Symons Downs, D. (2002a). Exercise
tress. Journal of Rational-Emotive and Cognitive-Behavior dependence: A systematic review. Psychology of Sport and
Therapy, 20, 33–47. doi: 10.1023/A:1015128904007 Exercise, 3, 89–123. doi: 10.1016/s1469-0292(00)00015-7

Journal of Behavioral Addictions 5(1), pp. 77–89 (2016) | 87

 Costa et al.

Hausenblas, H. A., & Symons Downs, D. (2002b). How much is Oliva, P., Costa, S., & Larcan, R. (2013). Physical self-concept and
too much? The development and validation of the exercise its relationship to exercise dependence symptoms in young
dependence scale. Psychology and Health, 17, 387–404. doi: regular physical exercisers. American Journal of Sports Sci-
10.1080/0887044022000004894 ence and Medicine, 1, 1–6. doi: 10.12691/ajssm-1-1-1
Hoek, H. W. (2006). Incidence, prevalence and mortality of anorexia Paulson, L. R., & Rutledge, P. C. (2014). Effects of perfection-
nervosa and other eating disorders. Current Opinion in Psychi- ism and exercise on disordered eating in college students.
atry, 19, 389–394. doi: 10.1097/01.yco.0000228759.95237.78 Eating Behaviors, 15, 116–119. doi: 10.1016/j.eatbeh.
Holmbeck, G. N. (1997). Toward terminological, conceptual, and 2013.11.005
statistical clarity in the study of mediators and moderators: Pole, R., Waller, D. A., Stewart, S. M., & Parkin‐Feigenbaum, L.
Examples from the child-clinical and pediatric psychology (1988). Parental caring versus overprotection in bulimia. In-
literatures. Journal of Consulting and Clinical Psychology, ternational Journal of Eating Disorders, 7, 601–606. doi:
65, 599–610. doi: 10.1037/0022-006x.65.4.599 10.1016/j.eatbeh.2013.11.005
Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in Polivy, J., & Herman, C. P. (2002). Causes of eating disorders.
covariance structure analysis: Conventional criteria versus new Annual Review of Psychology, 53, 187–213. doi: 10.1146/
alternatives. Structural Equation Modeling, 6, 1–55. doi: annurev.psych.53.100901.135103
10.1080/10705519909540118 Schreiber, K., & Hausenblas, H. A. (2015). The truth about
Klein, D. A., Bennett, A. S., Schebendach, J., Foltin, R. W., exercise addiction: Understanding the dark side of thinspira-
Devlin, M. J., & Walsh, B. T. (2004). Exercise “addiction” tion. Lanham, Maryland: Rowman & Littlefield.
in anorexia nervosa: Model development and pilot data. CNS Shafran, R., Cooper, Z., & Fairburn, C. G. (2002). Clinical perfec-
Spectrums, 9, 531–537. doi: 10.1017/S1092852900009627 tionism: A cognitive–behavioural analysis. Behaviour Research
Koenig, L. J., & Wasserman, E. L. (1995). Body image and dieting and Therapy, 40, 773–791. doi: 10.1016/s0005-7967(01)00059-6
failure in college men and women: Links between depression Shafran, R., & Mansell, W. (2001). Perfectionism and psychopa-
and eating problems. Sex Roles, 32, 225–249. doi: 10.1007/ thology: A review of research and treatment. Clinical Psycholo-
BF01544790 gy Review, 21, 879–906. doi: 10.1016/s0272-7358(00)00072-6
Kraemer, H. C., Stice, E., Kazdin, A., Offord, D., & Kupfer, D. (2001). Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and
How do risk factors work together? Mediators, moderators, and nonexperimental studies: New procedures and recommenda-
independent, overlapping, and proxy risk factors. American tions. Psychological Methods, 7, 422–445. doi: 10.1037/1082-
Journal of Psychiatry, 158, 848–856. doi: 10.1176/appi. 989x.7.4.422
ajp.158.6.848 Soenens, B., Elliot, A. J., Goossens, L., Vansteenkiste, M., Luyten,
Lewinsohn, P. M., Seeley, J. R., Moerk, K. C., & Striegel‐Moore, P., & Duriez, B. (2005). The intergenerational transmission of
R. H. (2002). Gender differences in eating disorder symptoms perfectionism: Parents’ psychological control as an intervening
in young adults. International Journal of Eating Disorders, 32, variable. Journal of Family Psychology, 19, 358–366. doi:
426–440. doi: 10.1002/eat.10103 10.1037/0893-3200.19.3.358
Little, T. D., Cunningham, W. A., Shahar, G., & Widaman, K. F. Soenens, B., Luyckx, K., Vansteenkiste, M., Luyten, P., Duriez,
(2002). To parcel or not to parcel: Exploring the question, B., & Goossens, L. (2008). Maladaptive perfectionism as an
weighing the merits. Structural Equation Modeling, 9, 151– intervening variable between psychological control and ado-
173. lescent depressive symptoms: A three-wave longitudinal study.
Lombardo, C. (2008). Adattamento italiano della multidimensional Journal of Family Psychology, 22, 465–474. doi: 10.1037/
perferctionism scale (mps) [Italian adaptation of the 0893-3200.22.3.465
multidimensional perfectionism scale (mps)]. Psicoterapia Soenens, B., & Vansteenkiste, M. (2010). A theoretical upgrade of
Cognitiva e Comportamentale, 14, 31–46. the concept of parental psychological control: Proposing new
Marsh, H. W., Hau, K. T., Balla , J. R., & Grayson, D. (1998). Is insights on the basis of self-determination theory. Develop-
more ever too much: The number of indicators per factor in mental Review, 30, 74–99. doi: 10.1016/j.dr.2009.11.001
confirmatory factor analysis. Multivariate Behavioral Re- Soenens, B., Vansteenkiste, M., Duriez, B., & Goossens, L. (2006).
search, 33, 181–220. In search of the sources of psychologically controlling parent-
May, A. L., Kim, J. Y., McHale, S. M., & Crouter, A. (2006). ing: The role of parental separation anxiety and parental
Parent–adolescent relationships and the development of weight maladaptive perfectionism. Journal of Research on Adoles-
concerns from early to late adolescence. International Journal cence, 16, 539–559. doi: 10.1111/j.1532-7795.2006.00507.x
of Eating Disorders, 39, 729–740. doi: 10.1002/eat.20285 Soenens, B., Vansteenkiste, M., & Luyten, P. (2010). Toward a
McCreary, D. R., & Sasse, D. K. (2000). An exploration of the domain-specific approach to the study of parental psychologi-
drive for muscularity in adolescent boys and girls. Journal of cal control: Distinguishing between dependency-oriented and
American College Health, 48, 297–304. doi: 10.1080/ achievement-oriented psychological control. Journal of Per-
07448480009596271 sonality, 78, 217–256.
Miller, K. J., & Mesagno, C. (2014). Personality traits and exercise Soenens, B., Vansteenkiste, M., Luyten, P., Duriez, B., &
dependence: Exploring the role of narcissism and perfection- Goossens, L. (2005). Maladaptive perfectionistic self-
ism. International Journal of Sport and Exercise Psychology, representations: The mediational link between psychological
12, 368–381. doi: 10.1080/1612197x.2014.932821 control and adjustment. Personality and Individual Differ-
Miller, S. R. (2012). I don’t want to get involved: Shyness, ences, 38, 487–498. doi: 10.1016/j.paid.2004.05.008
psychological control, and youth activities. Journal of Social Steinberg, L. (2001). We know some things: Parent–adolescent
and Personal Relationships, 7, 1–22. doi: 10.1177/ relationships in retrospect and prospect. Journal of Research
0265407512448266 on Adolescence, 11, 1–19. doi: 10.1111/1532-7795.00001

88 | Journal of Behavioral Addictions 5(1), pp. 77–89 (2016)

 Exercise dependence perfectionism parental control

Striegel‐Moore, R. H., Rosselli, F., Perrin, N., DeBar, L., Wilson, 1960 to 1979. International Journal of Eating Disorders, 16,
G. T., May, A., & Kraemer, H. C. (2009). Gender difference in 105–116. doi: 10.1002/1098-108x(199409)16:2 < 105::aid-
the prevalence of eating disorder symptoms. International Jour- eat2260160202 > 3.0.co;2-e
nal of Eating Disorders, 42, 471–474. doi: 10.1002/eat.20625 Vocks, S., Hechler, T., Rohrig, S., & Legenbauer, T. (2009).
Symons Downs, D., Hausenblas, H. A., & Nigg, C. R. (2004). Effects of a physical exercise session on state body image:
Factorial validity and psychometric examination of the The influence of pre-experimental body dissatisfaction and
exercise dependence scale-revised. Measurement in Physical concerns about weight and shape. Psychology and Health,
Education and Exercise Science, 8, 183–201. doi: 10.1207/ 24, 713–728. doi: 10.1080/08870440801998988
s15327841mpee0804_1 Zmijewski , C. F., & Howard, M. O. (2003). Exercise dependence
Vandereycken, W. (1994). Emergence of bulimia nervosa as a and attitudes toward eating among young adults. Eating Beha-
separate diagnostic entity: Review of the literature from viors, 4(10), 181–195.

Journal of Behavioral Addictions 5(1), pp. 77–89 (2016) | 89