Eating Disorders and Risk

Psychology of Sport & Exercise 5 (2004) 447–460
www.elsevier.com/locate/psychsport
Correlates of eating disorders risk among female figure
skates: a profile of adolescent competitors
5
Eva V. Monsma
a,Ã
, Robert M. Malina
b
a
Department of Physical Education, University of South Carolina, Columbia, SC 29208, USA
b
Tarleton State University, Box T-0010, Stephenville, TX 76402 USA
Received 2 July 2002; received in revised form 23 June 2003; accepted 3 July 2003
Abstract
Objectives: To present a profile of eating disorder risk among figure skaters by (a) comparing somato-
type and BMI as biological variables in explaining Eating Disorder Inventory (EDI) variance; (b) exam-
ining relationships among physical characteristics, physique-related perceptions, and psychological
indicators of eating disorders, and (c) determining if a composite of physical and psychological variables
could discriminate solo skaters from dance and pair skaters.
Design: Cross-sectional.
Methods: A battery of anthropometric dimensions was taken, and the Physical Self Description Ques-
tionnaire, Social Physique Anxiety Scale, and EDI were completed by 114 competitive female figure skat-
ers 12–22 yrs of age.
Results: Correlation analyses indicated that the BMI was the most robust biological variable for sub-
sequent analyses. Using hierarchical stepwise regression analysis to determine if physical self-perceptions
predicted EDI subscale score, age was forced into each analysis at the first step followed by the BMI at
step two. A composite of psychological variables (SPAS and select PSDQ subscales) were randomly
entered at step three. Beyond age and the BMI, each of the EDI subscales was predicted by at least one
psychological variable, with the SPAS predicting six of the eight subscales. Physical and psychological
variables accounted for 3% to 63% of the variance in the EDI subscale scores. Dancers and pair skaters
were older and less endomorphic, and reported higher Health scores, but lower Appearance scores com-
pared to solo skaters.
5
This manuscript was derived from a larger project titled ‘The Psychobiological Profile of Competitive Female
Figure Skaters’, funded by the Canadian Figure Skating Association (C.F.S.A.).
Ã
Corresponding author. Tel.: +1-803-777-1386; fax: +1-803-777-6250.
E-mail address: [email protected] (E.V. Monsma).
1469-0292/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S1469-0292(03)00038-4
Conclusion: Contextual, physical and psychological variables appear to be germane features of eating
disorder risk among competitive adolescent female figure skaters.
# 2003 Elsevier Ltd. All rights reserved.
Keywords: Eating disorder risk predictors; Physique; Body build; Psychological; Adolescence
The magnitude of eating disorder risk among female athletes is a topic of considerable
debate. Some studies suggest that athletes are at increased risk (Stoutjesdyk & Jevne, 1993;
Sundgot-Borgen, 1994), while others suggest that they may be buffered from eating disorders
(DiBartolo & Shaffer, 2002; Hausenblas & Mack, 1999). Correlates of eating disorder risk
among athletes have focused on sport environments which emphasize weight restriction such as
judo and lightweight rowing (Stoutjesdyk & Jevne, 1993), and which have a major aesthetic
component such as gymnastics, cheerleading, ballet, diving and figure skating (Brooks-Gunn,
Burrow, & Warren, 1988; DiBartolo & Shaffer, 2002; Reel & Gill, 1996; Sundgot-Borgen, 1994).
Although aesthetic sport participants generally self-report more drive for thinness and anorexic
symptomatology than non-aesthetic sport athletes (Hausenblas & Carron, 1999), research exam-
ining contributing factors specifically among figure skaters is sparse.
Research on skaters to date suggests a combination of environmental and individual factors
may be involved in the etiology of eating disorders is this population. Casual comments from
sport officials, coaches and/or peers (Gould, Jackson, & Finch, 1993) within the competitive
and/or practice environments can lead to negative self-perceptions associated with disordered
eating. Although skaters tend to be smaller, leaner and lighter (Brooks-Gunn, Burrow, & War-
ren, 1988) and are later maturing compared to non-athletes (Vadocz, Siegel, & Malina, 2002;
Ziegler, Hensley, et al., 1998), pressure to maintain or attain low body weight, and lack of
improvement associated with bodily changes of puberty are confirmed sources of stress among
figure skaters (Gould, Jackson, & Finch, 1993; Scanlan, Stein, & Ravizza, 1991). Skaters are
concerned about appearance, have lower than recommended daily caloric (energy) intake
(Rucinski, 1989; Ziegler, Khoo, et al., 1998), and often engage in caloric restriction (Ziegler,
Hensley, et al., 1998) and other behaviors to maintain or lose weight (Brooks-Gunn, Burrow, &
Warren, 1988). In a sample of 41 female pair and dance skaters 16–22 yr of age, EDI scores
resembled those of an eating disordered population more so than corresponding scores of non-
athlete female college students (Taylor & Ste.-Marie, 2001). These results also suggested that
skating with a male partner is a potential concomitant of eating disorders.
Behavioral context is a central theme in the developmental psychology literature. The good-
ness-of-fit hypothesis (Lerner, 1985) offers a potential explanation for the increased disordered
eating risk among skaters. According to this hypothesis, the valence of affect and behavior relies
on the extent to which an individual’s physical and psychological characteristics match contex-
tual demands, that is, the ‘goodness-of-fit’ between the athlete and her chosen sport. It is con-
ceivable that if the petite, linear physique of a skater matches the biomechanical (Harris, 1986)
and aesthetic demands of the sport, and if she has a positive self-concept beneficial to artistic
presentation and subjective evaluation, she may be buffered from negative affect and behavior.
In contrast, skaters who are unable to achieve or maintain characteristics demanded by the
E.V. Monsma, R.M. Malina / Psychology of Sport & Exercise 5 (2004) 447–460 448
sport may experience negative affect and perhaps resort to attempts at controlling physical size
by disordered eating behaviors. This contention is supported by clinical-based explanations of
eating disorders suggesting a relationship between self-restricted eating and perceived control
(Rezek & Leary, 1991). Restricted eating among aesthetic sport athletes may in part be a
response to a lack of control over performance outcome primarily controlled by judges.
A self-presentational understanding of eating disorders has been advocated (Hausenblas &
Mack, 1999; Leary, Tchividjian, & Kraxberger, 1994) and is consistent with the ‘goodness-of-fit
hypothesis.’ Impression management theory suggests that Social Physique Anxiety (SPA) is
involved in the selective presentation of the self in order to create desired impressions and to
avoid those undesired (Leary, 1992). Accordingly, skaters who perceive that they do not meet
the physical characteristics preferred by sport officials may experience heightened SPA—a liab-
ility in a sport where success is contingent on judges’ evaluations of self-presentation. Social
physique anxiety and eating disorder risk often develop early in adolescence (Crocker, Synder,
Kowalski, & Hoar, 2001; Thompson & Chad, 2002). Figure skating is an early entry sport
where specialization often occurs prior to 8 yr of age and where physique and appearance are
evaluated regularly. Thus, a positive self-concept is essential particularly when competition and
pubertal transitions coincide. Researchers have suggested that changes associated with growth
and maturation negatively impact self-concept (Brooks-Gunn, 1988; Marsh, 1996, 1998). Thus,
in addition to SPA, age-related physical characteristics and self-concept are warranted when
considering the etiology of eating disorders.
Physical characteristics (size, physique, proportions, composition) should be considered in
evaluating correlates of disordered eating risk in athletes. Variance associated with physical
characteristics and with specific psychological variables, and perhaps their interactions may bet-
ter explain the determinants of risk for disordered eating. The most frequently used physical
characteristic is the body mass index [BMI: weight (kg)/height (m
2
)], which is often used as an
indicator of fatness. However, it is also related to fat-free mass and is more appropriately an
indicator of heaviness. It may have limited utility with athletes who are generally more muscular
and less fat than the general population (Malina, Bouchard, & Bar-Or, 2004). Athletes may
have an elevated BMI because they are heavier and leaner.
An alternative for assessing the contribution of physical characteristics to the risk of eating
disorders is an objective estimate of physique, or body build. Somatotype is a quantitative
description of physique which has a long tradition of use in studies of athletes (Carter & Heath,
1991). An individual’s somatotype is a composite of three components: endomorphy (relative
fatness), mesomorphy (relative muscularity) and ectomorphy (relative linearity). Somatotype
may be a more robust correlate of eating disorder risk since it includes specific features of phys-
ique and also permits evaluations of intra-individual variation, that is, individuals may have the
same BMI but can differ in somatotype. Somatotype changes are generally small during child-
hood, but are variable during puberty and the adolescent growth spurt due to individual differ-
ences in timing and tempo of these processes (Malina et al., 2004).
The purpose of this study was twofold. First, it sought to examine potential physical, psycho-
logical and contextual variables thought to contribute to the etiology of disordered eating, and
second, it attempted to determine if selected physical and psychological characteristics discrimi-
nate skaters who skated with male partners from solo skaters. Several hypotheses were gen-
erated in the context of the two purposes. Somatotype and the BMI are related to indices of
449 E.V. Monsma, R.M. Malina / Psychology of Sport & Exercise 5 (2004) 447–460
eating disorder. Somatotype is a stronger physical correlate of eating disorder indices than the
BMI. Social physique anxiety and perceptions associated with physical characteristics are salient
predictors of disordered eating indices. And, skaters with male partners differ in physique and
psychological characteristics from solo skaters.
Methodology
Participants
Participants were 114 female figure skaters, 12.8–22.3 yr of age (mean age ¼ 16:2 Æ2:0 yr).
They were participants in four United States Figure Skating Association (USFSA) clubs
(n ¼ 25), and seven Canadian Figure Skating Association (CFSA) clubs (n ¼ 89). Purposive
sampling was used to recruit skaters across three disciplines (solo: n ¼ 74; dance: n ¼ 22, pairs:
n ¼ 18). Participants were part of a larger study of physical and psychological characteristics of
figure skaters (Vadocz, 1999). The response rate for participants completing all of the psycho-
logical inventories was 71%. Status quo and retrospective menarcheal data for the total sample
of skaters (n ¼ 159) are reported in Monsma, Malina & Feltz (2003), and variation in physical
and psychological variation characteristics by maturational status and timing are presented in
Vadocz et al. (2002).
Variables and measurement procedures
General procedures
Officials from the skating clubs were contacted by phone to request permission for contacting
the figure skaters. Participation and consent were then solicited from parents and athletes at an
information meeting. All subjects and/or their parents gave informed consent. The study was
approved by the University Committee for Research Involving Human Subjects at Michigan
State University. Consenting skaters were given a questionnaire package and a tentative time
for anthropometry was scheduled. The questionnaire package included a form on background
and training history in figure skating and the forms for assessing eating disorder risk, social
physique anxiety, and physical self-description scales. Skaters were instructed to complete the
package and return it to the primary investigator at the time scheduled for anthropometry,
which was scheduled around practice schedules within one month of the initial request for par-
ticipation. The menarcheal status of the skaters was ascertained at interview. Consistent with
Hausenblas and Carron’s (1999) recommendations, all data were collected during January
through March, the peak figure skating training season.
Psychological measures
Risk of eating disorder was assessed with the Eating Disorder Inventory (EDI, Garner, Olm-
stead, & Polivy, 1983). The 64-item self-report inventory consists of eight subscales: Drive for
Thinness, Body Dissatisfaction, Bulimia, Perfectionism, Maturity Fears, Ineffectiveness Inter-
E.V. Monsma, R.M. Malina / Psychology of Sport & Exercise 5 (2004) 447–460 450
personal Distrust, and Introceptive Awareness. Scores range from lowest to highest where
1 ¼ ‘never’, 2 ¼ ‘rarely’, 3 ¼ ‘sometimes’, 4 ¼ ‘often’, 5 ¼ ‘usually’, and 6 ¼ ‘always’. Scoring
the EDI requires summing the responses after condensing the 6-point Likert scale to a 4-point
scale where scores of 1–3 are recoded as 0, 4 is recoded as 1, 5 is recoded as 2, and 6 is recoded
as 3. Individuals scoring 15 and 10 on the summed scores for the Drive for Thinness and
Body Dissatisfaction scales, respectively, are considered at risk for developing an eating disorder
(Garner & Olmstead, 1984).
The Physical Self-Description Questionnaire (PSDQ, Marsh, 1996) was used as a measure of
self-concept. It is a 70-item self-report scale measuring nine components of Physical Self-Con-
cept: Appearance, Coordination, Endurance, Flexibility, Strength, Health, Physical Activity,
Body Fat, and Sport Competence. The scale also provides two global components: Global
Physical Self-Concept and Self-Esteem. Each item is a declarative statement and the participant
responds on a 6-point true-false response scale. It has been used with youth, 12–18 yr of age,
and has been recommended as appropriate for older subjects (Marsh, Richards, Johnson,
Roche, & Tremayne, 1994).
The 9-item Social Physique Anxiety Scale (SPAS) recommended by Martin, Rejeski, Leary,
McAuley, and Bane (1997) was used to assess the anxiety that individuals experience in
response to others’ evaluations of their physique. This modification omits questions 2, 5, and 11
from the original 12-item questionnaire due to conceptual and empirical weaknesses. Each item
is scored on a 5-point scale where 1 ¼ ‘not at all’, 2 ¼ ‘slightly’, 3 ¼ ‘moderately’, 4 ¼ ‘very’ (4),
5 ¼ ‘extremely’. Items 1 and 8 are reverse scored.
Anthropometry and physique
Anthropometric dimensions needed to estimate somatotype with the Heath-Carter protocol
(Carter & Heath, 1991) were taken by the primary researcher following the methods described
by Lohman, Roche, and Martonell (1988): weight (kg), height (cm), flexed arm circumference
(cm), calf circumference (cm), biepicondylar breadth of the humerus (cm), bicondylar breadth of
the femur (cm), and four skinfolds (mm)—triceps, subscapular, supraspinale and medial calf.
Multiple circumferences, breadth and skinfold measurements were taken (2, 2, and 3 times,
respectively), and the averages of the measures were used in the analyses. Inter-rater reliability
between an experienced anthropometrist and the primary investigator, and intra-rater reliability
were established on 13 subjects who were not part of the study sample. The technical error of
measurement (TEM, Malina, 1995) was used as the indicator of measurement variability. TEM
¼
p
ð
P
d
2
Þ=2n is the square root of the sum of squared differences of replicate measurements
(
P
d
2
) divided by twice the number of pairs (2n). Inter- and intra-observer errors for the anthro-
pometric dimensions were within acceptable ranges for studies of young athletes, 0.01–0.64
(Malina, 1995).
The BMI was calculated as weight (kg) divided by height squared (m
2
). Somatotype was esti-
mated with the Heath–Carter algorithms:
Endomorphy ¼ À0:7182 þ 0:1451ðXÞ À 0:00068ðX
2
Þ þ 0:0000014ðX
3
Þ
451 E.V. Monsma, R.M. Malina / Psychology of Sport & Exercise 5 (2004) 447–460
T
a
b
l
e
1
D
e
s
c
r
i
p
t
i
v
e
s
t
a
t
i
s
t
i
c
s
f
o
r
b
i
o
l
o
g
i
c
a
l
c
h
a
r
a
c
t
e
r
i
s
t
i
c
s
a
n
d
p
s
y
c
h
o
l
o
g
i
c
a
l
v
a
r
i
a
b
l
e
s
a
n
d
p
a
r
t
i
a
l
c
o
r
r
e
l
a
t
i
o
n
s
(
a
g
e
c
o
n
s
t
a
n
t
)
w
i
t
h
E
D
I
s
u
b
s
c
a
l
e
s
c
o
r
e
s
a
m
o
n
g
fi
g
u
r
e
s
k
a
t
e
r
s
(
n
¼
1
1
4
)
E
D
I
s
u
b
s
c
a
l
e
s
a
D
e
s
c
r
i
p
t
i
v
e
s
t
a
t
i
s
t
i
c
s
M
S
D
1
2
3
4
5
6
7
8
M
5
.
1
7
.
2
1
.
8
5
.
4
6
.
2
2
.
3
2
.
8
2
.
3
S
D
6
.
1
7
.
5
3
.
7
4
.
1
4
.
1
0
.
7
0
.
8
0
.
8
B
i
o
l
o
g
i
c
a
l
v
a
r
i
a
b
l
e
s
A
g
e
(
y
r
)
1
5
.
9
2
.
2
0
.
3
7
Ã
Ã
Ã
0
.
4
2
Ã
Ã
Ã
0
.
2
3
Ã
Ã
À
0
.
0
5
0
.
1
1
0
.
3
0
Ã
Ã
Ã
0
.
0
7
0
.
2
9
Ã
Ã
Ã
H
e
i
g
h
t
(
c
m
)
1
5
8
.
3
7
.
0
0
.
3
5
Ã
Ã
Ã
0
.
4
1
Ã
Ã
Ã
0
.
1
8
Ã
0
.
1
0
0
.
0
1
0
.
3
5
Ã
Ã
Ã
0
.
1
4
0
.
1
9
Ã
W
e
i
g
h
t
(
k
g
)
5
0
.
8
8
.
6
0
.
4
1
Ã
Ã
Ã
0
.
5
1
Ã
Ã
Ã
0
.
1
8
Ã
0
.
0
8
À
0
.
0
5
0
.
3
8
Ã
Ã
Ã
0
.
1
0
0
.
2
3
Ã
Ã
B
M
I
(
k
g
/
m
2
)
2
0
.
7
2
.
4
0
.
4
0
Ã
Ã
Ã
0
.
5
4
Ã
Ã
Ã
0
.
1
8
Ã
0
.
0
5
À
0
.
0
6
0
.
2
9
Ã
Ã
À
0
.
0
6
0
.
1
9
Ã
E
n
d
o
m
o
r
p
h
y
a
3
.
4
1
.
0
0
.
0
0
0
.
1
3
0
.
0
3
À
0
.
0
4
À
0
.
0
1
0
.
0
5
À
0
.
1
2
À
0
.
1
4
M
e
s
o
m
o
r
p
h
y
a
3
.
7
0
.
9
0
.
0
1
À
0
.
1
0
0
.
0
1
0
.
0
1
0
.
1
6
À
0
.
1
8
Ã
À
0
.
1
3
À
0
.
2
0
Ã
E
c
t
o
m
o
r
p
h
y
a
2
.
9
1
.
0
À
0
.
1
5
0
.
1
9
À
0
.
0
1
0
.
0
0
0
.
0
0
À
0
.
1
8
Ã
À
0
.
1
0
À
0
.
2
5
Ã
Ã
P
s
y
c
h
o
l
o
g
i
c
a
l
v
a
r
i
a
b
l
e
s
P
S
D
Q
A
p
p
e
a
r
a
n
c
e
2
6
.
9
5
.
3
À
0
.
2
8
Ã
Ã
À
0
.
4
9
Ã
Ã
Ã
À
0
.
2
8
Ã
Ã
À
0
.
1
4
À
0
.
0
8
À
0
.
5
8
Ã
Ã
Ã
À
0
.
3
5
Ã
Ã
Ã
À
0
.
4
1
Ã
Ã
Ã
B
o
d
y
f
a
t
2
8
.
2
8
.
0
À
0
.
6
5
Ã
Ã
Ã
À
0
.
7
7
Ã
Ã
Ã
À
0
.
3
4
Ã
Ã
Ã
À
0
.
2
1
Ã
Ã
À
0
.
1
0
À
0
.
5
4
Ã
Ã
Ã
À
0
.
0
8
À
0
.
5
0
Ã
Ã
Ã
H
e
a
l
t
h
3
7
.
4
6
.
9
À
0
.
0
8
À
0
.
0
4
À
0
.
2
7
Ã
Ã
À
0
.
0
9
À
0
.
2
3
Ã
Ã
À
0
.
1
8
Ã
À
0
.
1
3
À
0
.
2
9
Ã
Ã
G
l
o
b
a
l
p
h
y
s
i
c
a
l
s
e
l
f
-
c
o
n
c
e
p
t
2
7
.
7
6
.
5
À
0
.
5
5
Ã
Ã
Ã
À
0
.
6
9
Ã
Ã
Ã
À
0
.
3
2
Ã
Ã
Ã
À
0
.
1
5
0
.
0
1
À
0
.
6
8
Ã
Ã
Ã
À
0
.
2
4
Ã
Ã
À
0
.
5
2
Ã
Ã
Ã
S
e
l
f
-
e
s
t
e
e
m
4
1
.
9
5
.
3
À
0
.
4
4
Ã
Ã
Ã
À
0
.
4
8
Ã
Ã
Ã
À
0
.
3
2
Ã
Ã
Ã
À
0
.
2
5
Ã
Ã
À
0
.
0
2
À
0
.
7
2
Ã
Ã
Ã
À
0
.
3
2
Ã
Ã
Ã
À
0
.
4
6
Ã
Ã
Ã
S
P
A
2
4
.
7
6
.
7
0
.
6
0
Ã
Ã
Ã
0
.
5
6
Ã
Ã
Ã
0
.
3
6
Ã
Ã
Ã
0
.
3
5
Ã
Ã
Ã
0
.
1
8
Ã
0
.
4
8
Ã
Ã
Ã
0
.
1
7
0
.
5
3
Ã
Ã
Ã
a
C
o
n
t
r
o
l
l
i
n
g
f
o
r
a
g
e
a
n
d
t
w
o
s
o
m
a
t
o
t
y
p
e
c
o
m
p
o
n
e
n
t
s
(
C
r
e
s
s
i
e
,
W
i
t
h
e
r
s
,
&
G
r
a
i
g
,
1
9
8
6
)
;
1
=
d
r
i
v
e
f
o
r
t
h
i
n
n
e
s
s
,
2
=
b
o
d
y
d
i
s
s
a
t
i
s
f
a
c
t
i
o
n
,
3
=
b
u
l
i
m
i
a
,
4
=
m
a
t
u
r
i
t
y
f
e
a
r
s
,
5
=
p
e
r
f
e
c
t
i
o
n
i
s
m
,
6
=
i
n
e
ff
e
c
t
i
v
e
n
s
s
,
7
=
i
n
t
e
r
p
e
r
s
o
n
a
l
d
i
s
t
r
u
s
t
,
8
=
i
n
t
r
o
c
e
p
t
i
v
e
a
w
a
r
e
n
e
s
s
.
T
h
i
r
d
o
r
d
e
r
p
a
r
t
i
a
l
c
o
r
-
r
e
l
a
t
i
o
n
s
,
h
o
l
d
i
n
g
a
g
e
,
a
n
d
t
h
e
o
t
h
e
r
t
w
o
s
o
m
a
t
o
t
y
p
e
c
o
m
p
o
n
e
n
t
s
c
o
n
s
t
a
n
t
.
Ã
P
<
0
:
0
5
;
Ã
Ã
P
<
0
:
0
1
;
Ã
Ã
Ã
P
<
0
:
0
0
1
.
E.V. Monsma, R.M. Malina / Psychology of Sport & Exercise 5 (2004) 447–460 452
where X is the sum of the triceps, subscapular and supraspinale skinfolds, adjusted for stature;
Mesomorphy ¼ ð0:858 biepicondylar þ 0:601 bicondylar
þ 0:188 corrected arm circumference
þ 0:161 corrected calf circumferenceÞ Àðstature  0:131Þ þ 4:50
where corrected arm and calf circumferences are the respective limb circumferences minus the
triceps and medial calf skinfolds, respectively;
Ectomorphy ¼ HWR Â 0:732 À 28:58
where HWR ¼ stature
3
p
weight. If HWR < 40:75 but >38.25, ectomorphy ¼ HWRÂ
0:463 À17:63. If HWR 38:25, a rating of 0.1 is assigned (Carter & Heath, 1991).
Each component has a numerical score ranging from a low value 0.1 and in theory no upper
limit, although most scores are below seven. The component ratings are always presented in the
same order: endomorphy, mesomorphy, ectomorphy. A somatotype rating of 2-5-2 thus indi-
cates an individual high in mesomorphy with rather slight development of endomorphy and
ectomorphy.
Results
Descriptive statistics and correlational analyses
Cronbach (1951) alpha coefficients ranged from 0.70 to 0.94 for the EDI scales, and from
0.81 to 0.94 for the PSDQ subscales in the figure skaters. It was 0.91 for the SPAS. BMIs
ranged from 14.9 to 24.3 kg/m
2
and 25% of the skaters (n ¼ 28) in the sample were pre-
menarcheal. Forty-three (38%) of the skaters met the eating disorder risk criterion for Drive for
Thinness (score >15) and 62 (54%) met the Body Dissatisfaction eating disorder risk criterion
(score >10), whereas 13 (8%) met both the Drive for Thinness and Body Dissatisfaction criteria.
Screening for multicollinearity of all independent and dependent variables showed no correla-
tions > 0.90 (Tabachnick & Fidell, 1996). Evaluations for normal distribution indicated the
need to transform several variables to reduce skewed data. Due to moderate negative skews in
Health, Global Physical Self-Concept, Appearance and Self-Esteem, reflect and square root
transformations were used. Square root transformations were used for scores on Bulimia and
Body Fat because they were somewhat positively skewed (Tabachnick & Fidell, 1996).
Correlations between age and EDI subscale scores were, with two exceptions, low to moder-
ate and positive. Hence, age was controlled in subsequent analyses. Means and standard devia-
tions for all variables, and partial correlations with age held constant, between EDI subscales
and the physical and psychological variables are summarized in Table 1. Correlations between
physical and psychological characteristics and EDI subscales were variable but in the expected
directions. Correlations between height, weight and the BMI, and Drive for Thinness, Body
Dissatisfaction, Bulimia, Ineffectiveness and Introceptive Awareness were positive and moder-
ate. Mesomorphy and ectomorphy were correlated with Ineffectiveness and Introceptive Aware-
ness, and the relationships were negative and low. The BMI was a stronger correlate of EDI
subscales than the somatotype components. Significant low to moderate positive correlations
were found between BMI and five of the EDI subscales whereas only mesomorphy and
453 E.V. Monsma, R.M. Malina / Psychology of Sport & Exercise 5 (2004) 447–460
ectomorphy components of somatotype showed significant negative correlations with two of the
social adjustment subscales (Ineffectiveness and Introceptive Awareness). Consequently, the
BMI was used as the physical (biological) control variable in subsequent analyses. Consistent
Table 2
Results of the regression analysis for variables predicting EDI subscale scores
Variables Adjusted Standard
r R
2
R
2
Increment Beta t
Drive for thinness
Age 0.25 0.06 0.06 0.06 À0.03 À0.37
BMI 0.35 0.12 0.11 0.05 0.01 0.14
Body fat 0.63 0.40 0.38 0.24 À0.41 À4.30
ÃÃÃ
SPA 0.70 0.48 0.47 0.09 0.37 4.25
ÃÃÃ
Body dissatisfaction
Age 0.36 0.13 0.13 0.13 0.02 0.35
BMI 0.56 0.33 0.32 0.20 0.17 2.19
Ã
Body fat 0.75 0.57 0.55 0.24 À0.40 À5.00
ÃÃÃ
Global physical self-concept 0.80 0.64 0.63 0.08 0.36 4.76
ÃÃÃ
Bulimia
Age 0.12 0.02 0.01 0.02 0.02 0.16
BMI 0.14 0.02 0.00 0.01 À0.06 À0.56
SPA 0.40 0.15 0.13 0.13 0.41 4.18
ÃÃÃ
Perfectionism
Age 0.01 0.00 À0.01 0.00 0.00 0.02
BMI 0.09 0.01 À0.01 0.01 À0.19 À1.69
SPA 0.24 0.06 0.03 0.05 0.25 2.40
Ã
Ineffectiveness
Age 0.02 0.00 À0.01 0.00 À0.12 À1.18
BMI 0.08 0.01 À0.01 0.01 À0.05 À0.48
SPA 0.36 0.13 0.11 0.13 0.40 4.00
ÃÃÃ
Maturity fears
Age 0.25 0.06 0.06 0.06 À0.27 À2.81
ÃÃ
BMI 0.26 0.07 0.05 0.00 À0.10 À0.90
Self-esteem 0.40 0.16 0.14 0.10 0.32 3.26
ÃÃ
SPA 0.45 0.20 0.18 0.04 0.34 3.18
ÃÃ
Body fat 0.50 0.25 0.21 0.04 0.31 2.48
Ã
Interpersonal distrust
Age 0.05 0.00 À0.01 0.00 À0.04 À0.40
BMI 0.09 0.01 À0.01 0.01 À0.18 À1.70
Self-esteem 0.39 0.15 0.13 0.15 0.26 2.27
Ã
Appearance 0.43 0.18 0.15 0.03 0.22 2.00
Ã
Introceptive awareness
Age 0.15 0.02 0.01 0.02 À0.02 À0.22
BMI 0.18 0.03 0.02 0.02 À0.20 À1.89
Global physical self-concept 0.49 0.24 0.22 0.20 0.27 2.50
ÃÃ
SPA 0.55 0.30 0.28 0.07 0.26 2.43
Ã
Body fat 0.57 0.33 0.30 0.03 À0.25 À2.12
Ã
Ã
P < 0:05;
ÃÃ
P < 0:01;
ÃÃÃ
P < 0:001.
E.V. Monsma, R.M. Malina / Psychology of Sport & Exercise 5 (2004) 447–460 454
with the hypotheses, correlations between PSDQ and EDI subscales were negative and ranged
from low to high, whereas correlations between SPA and each EDI subscale were positive and
moderate. Each of the selected psychological variables showed significant correlations with at
least four of the eight EDI subscales were thus, entered as predictors in the regression analyses.
Regression analyses
Eight separate regression analyses, one for each EDI subscale, were conducted. Age was
forced into each analysis at step one and the BMI was forced into the equation at step two.
Next, the following psychological variables were entered using stepwise regression analysis at
step 3: Appearance, Body Fat, Health, Global Physical Self-Concept, and Self-Esteem from the
PSDQ and SPA. This quasi-hierarchical approach was utilized to control for age and body size
before considering the amount of variance accounted for by psychological variables.
Results of the regression equation are presented in Table 2. The variance in Drive for Thin-
ness was accounted for by Body Fat (24%) and SPA (9%) accounting for 47% of the total vari-
ance Fð4; 113Þ ¼ 25:61, P < 0:001. In addition to the BMI (20%), Body Dissatisfaction was
predicted by Body Fat (24%) and Global Physical Self-Concept (8%), accounting for 63% of the
total variance Fð4; 113Þ ¼ 48:60, P < 0:001. SPA was the only significant predictor of Bulimia,
Fð3; 113Þ ¼ 6:68, P < 0:001, Perfectionism, Fð3; 113Þ ¼ 2:25, P < 0:05, and Ineffectiveness,
Fð3; 113Þ ¼ 5:57, P ¼ 0:001, accounting for13%, 3% and 11% of variances, respectively.
Maturity Fears was predicted by age (6%), Self-Esteem (10%), SPA (4%) and Body Fat (4%),
accounting for 21% of the total variance Fð5; 113Þ ¼ 7:08, P < 0:001. Self-Esteem (15%) and
Appearance (3%) Fð3; 113Þ ¼ 6:10, P < 0:001, accounted for 15 % of the variance in Inter-
personal Distrust, and Global Physical Self-Concept (20%), SPA (7%) and Body Fat (3%),
Fð4; 113Þ ¼ 10:61, P < 0:001, accounted for 30% of the variance in Introceptive Awareness.
Discriminant function analysis
A discriminant function analysis was performed to determine if physical and psychological
variables discriminated between skaters skating with a partner (dance and pair skaters: n ¼ 40)
and those skating solo (solo skaters: n ¼ 74). Age, the BMI, somatotype, the PSDQ subscales of
Health, Appearance, Global Physical Self-Concept, Self-Esteem, Body Fat and SPA were varia-
bles included in the analysis (see Table 3). Since the homogeneity of variance was significant for
Table 3
Results of the discriminant function analysis for discipline
Variable Discipline Wilks’ Lambda Standard
function
F Adj. M SD Adj. M SD
Solo (n ¼ 74) Dance and pairs
(n ¼ 40)
1. Health 0.86 0.50 18.6 0.9 0.3 1.1 0.2
2. Endomorphy 0.80 0.63 13.8 3.6 1.1 2.9 0.8
3. Age 0.76 À0.39 11.5 15.6 2.2 16.5 2.4
4. Appearance 0.73 À0.61 1.0 1.0 0.2 0.9 0.3
455 E.V. Monsma, R.M. Malina / Psychology of Sport & Exercise 5 (2004) 447–460
discipline, Box’s M ¼ 28:4, Fð10Þ ¼ 2:7, P < 0:01, separate covariance matrices were used for
the analysis (Tabachnick & Fidell, 1996).
The analysis discriminated across discipline. Group centroids were À0.440 for solo skaters
and 0.815 for the combined sample of dance and pair skaters. Age, endomorphy, satisfaction
with Health and Appearance correctly classified 79% of the skaters. Dancers and pair skaters
were older and less endomorphic, and reported higher Health scores, but lower Appearance
scores compared to solo skaters (see Table 3).
Discussion
From a contextual perspective, the goal of this study was to generate a profile of figure skat-
ers at risk for disordered eating. Collectively, eating disorder risk in this sample of skaters was
somewhat similar to previous investigations. Although as a group, the mean EDI subscale
scores (Table 1) were lower than previous studies of athletes (Hausenblas & Mack, 1999; Taylor
& Ste.-Marie, 2001), the percentage of skaters meeting the criteria for eating disorder risk was
similar to other reports of figure skaters (Taylor & Ste.-Marie, 1998) and gymnasts (Petrie,
1993).
Beyond the possibility of inherent social desirability concerns in eating disorder research, the
broad age range of the sample (12–22 yr) may explain in part the lower mean scores. Changes
in body form, distortion of body image and desire to become thinner accompany sexual matu-
ration (Attie & Brooks-Gunn, 1989). Since 25% of the skaters were pre-menarcheal, their con-
cerns for factors associated with the risk of disordered eating may not have been strong. The
correlations between physical variables, physical self-perceptions and EDI subscales in the figure
skaters have several implications. The majority of EDI subscale scores increased with age and
with changes in physical characteristics associated with growth and maturation. In addition to
supporting the validity of the psychological instruments used in this investigation, this pattern
of correlations corroborates assertions that eating disorders may be a function of changing
physical characteristics (Attie & Brooks-Gunn, 1989; Hausenblas & Carron, 1999; Killen et al.,
1992).
Controlling for the BMI as an indicator of body size and composition is often advocated in
the eating disorder literature (Hausenblas & Mack, 1999; Petrie, 1996). The present study is
unique since it also considered an objective measure of physique (somatotype) as a physical cor-
relate. Contrary to what was expected, the BMI was the most robust correlate of EDI subscales
and was thus used as a physical control in subsequent analyses. Surprisingly, somatotype com-
ponents were not stronger correlates than the BMI. This may be related to the rather narrow
range of variation in somatotype that characterizes athletes in many sports compared to the
general population of adolescent girls (Carter & Heath, 1991). The lack of shoulder and hip
dimensions in the estimate of mesomorphy in the Heath-Carter protocol may contribute to the
lack of stronger correlations. Shoulder and hip dimensions change absolutely and relative to
each other during the growth spurt and sexual maturity (Malina et al., 2004), these changing
relationships are often a concern for adolescent girls. Female athletes (and often their coaches)
are commonly concerned about body weight and heaviness per se, and it is possible that their
overall morphology as reflected in somatotype is not a primary concern, especially in sports like
E.V. Monsma, R.M. Malina / Psychology of Sport & Exercise 5 (2004) 447–460 456
figure skating, artistic gymnastics and ballet where the range of variation in physique is often
limited. Thus, weight, height and the BMI are physical characteristics that are more strongly
subject to perception than somatotype. This notion is consistent with the literature suggesting
that perceived heaviness is the important factor in the risk of disordered eating (e.g., Ackard &
Peterson, 2001; Cooley & Toray, 2001).
It is important to note, however, that BMIs are often based on self-reported rather than
measured heights and weights in eating disorder studies involving athletes (Diehl, Johnson,
Rogers, & Petrie, 1998; Hausenblas & Mack, 1999; Petrie, 1996). This may be a major limi-
tation in studies of adolescents. Social desirability may influence underreporting weight and the
focus on size in some sports may influence over-estimating height. Further, individual differ-
ences in the rate of growth during adolescence may influence reported heights and weights; ado-
lescents who are in the midst of their growth spurt may not realize the magnitude of recent
growth (Himes & Faricy, 2001).
After controlling for age and the BMI, SPA was related to the EDI accounting for 4% to
13% of the variance in five of the eight subscales. Consistent with our hypotheses, this finding
suggests that disordered eating may in part, be a response to physique anxiety stemming from
the subjective evaluative nature of the sport. The positive correlation between SPA and
Maturity Fears further suggests that SPA may be a greater concern during physical develop-
ment. This finding should help guide longitudinal investigations towards better understanding
the possible a priori contribution of SPA to eating disorder risk in relation to physical changes
occurring with growth and maturation.
The relationship between SPA and Bulimia, and between SPA and two EDI social adjustment
scales (Ineffectiveness and Introceptive Awareness) should raise concerns for practitioners. Since
SPA may develop in early adolescence (Crocker et al., 2001; Vadocz, 1999) and eating patholo-
gies stem in part from physique-related concerns (Diehl et al., 1998; Hausenblas & Mack, 1999;
Leary, Tchividjian, & Kraxberger, 1994), early adolescent girls with concerns for self-presen-
tation may be motivated to engage in problematic weight control behaviors (e.g., restricting and
purging). Weight control behaviors may be an attempt to regain control over their changing
physique or performance environment that is largely controlled by judges (Rezek & Leary,
1991). Clearly, more research is needed concerning the prevalence of SPA among younger girls.
Practitioners may consider directing aesthetic activity participants to exert control over other
areas than eating.
A consistent trend of negative correlations between the EDI subscales and physical self-per-
ceptions indicated that skaters less satisfied with their physical characteristics reported higher
EDI scores. Aligned with previous research and our hypothesis, physical self-perceptions appear
to be salient features of eating disorder risk (Cooley & Toray, 2001; Diehl et al., 1998). Physical
self-perceptions explained a significant portion of the variance in each EDI subscale after con-
trolling for age and body size. Satisfaction with Body Fat predicted both Drive for Thinness
and Body Dissatisfaction. In addition, skaters who reported lower Global Physical Self-Concept
reported higher Body Dissatisfaction.
From a practical standpoint, the results suggest that risk of developing eating disorders
increases with the BMI. This may apply more so in athletes who are concerned about presenting
themselves in front of others and who may have negative perceptions of their physical charac-
teristics. In addition to employing cognitive restructuring strategies aimed at improving overall
457 E.V. Monsma, R.M. Malina / Psychology of Sport & Exercise 5 (2004) 447–460
physical self-perceptions and reducing social physique anxiety, practitioners should perhaps
focus on perceptions of weight because figure skaters as a group are not overweight compared
to non-athletes of the same age (Monsma & Malina, in press).
Finally, the role of psychobiological variables in discriminating across discipline was con-
sidered. In line with Lerner’s (1985) matching hypothesis and with the results of Taylor and Ste.-
Marie (2001), pair and dance skaters were expected to differ from solo skaters. Dancers and pair
skaters were older, less endomorphic, and reported higher Health, but slightly lower Appearance
scores than solo skaters (Table 3). It is possible that dancers and pair skaters reported feeling
healthier because they were less endomorphic, that is, less fat. It is also possible that they reported
less satisfaction with their appearance because they were older and perhaps recognized changes in
physique associated with age during adolescence. Physical changes associated with growth and
maturation are recognized sources of stress among figure skaters (Gould, Jackson, & Finch, 1993;
Scanlan, Stein, & Ravizza, 1991) and would be considered liabilities for skaters who are chosen to
skate with male partners largely because of their petite physique.
This investigation highlights the importance of considering the interaction of psychological,
biological and contextual variables in attempts to explain the etiology of eating disorders.
Future investigations should consider information on male partners and the potential mediating
effects of SPA and self-concept in the development of disordered eating. Maintaining or improv-
ing physical self-perceptions may be an important foundation in the prevention of eating dis-
orders, especially with skaters who skate with male partners.
References
Ackard, D. M., & Peterson, C. B. (2001). Association between puberty and disordered eating, body image and other
psychological variables. International Journal of Eating Disorders, 29, 187–194.
Attie, I., & Brooks-Gunn, J. (1989). Development of eating problems in adolescent girls: a longitudinal study. Devel-
opmental Psychology, 25, 70–79.
Brooks-Gunn, J. (1988). Antecedents and consequences of variations in girls’ maturational timing. Journal of Ado-
lescent Health Care, 9, 365–373.
Brooks-Gunn, J., Burrow, C., & Warren, M. P. (1988). Attitudes toward eating and body weight in different groups
of female adolescent athletes. International Journal of Eating Disorders, 7, 749–757.
Carter, J. E. L., & Heath, B. H. (1991). Somatotyping: development and applications. New York: Cambridge Univer-
sity Press.
Cooley, E., & Toray, T. (2001). Body image and personality predictors of eating disorder symptoms during the col-
lege years. International Journal of Eating Disorders, 30, 28–36.
Cressie, N. A. C., Withers, R. T., & Graig, N. P. (1986). The statistical analysis of somatotype data. Yearbook of
Physical Anthropology, 29, 197–208.
Crocker, P. R. E., Synder, J., Kowalski, K. C., & Hoar, S. (2001). Don’t let me be fat or physically incompetent! The
relationship between physical self-concept and social physique anxiety in Canadian high performance female ado-
lescent athletes. AVANTE, 6, 16–23.
Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrika, 16, 297–334.
DiBartolo, P. M., & Shaffer, C. (2002). A comparison of female college athletes and non-athlete: eating disorder
symptomatology and psychological well-being. Journal of Sport and Exercise Psychology, 24, 33–41.
Diehl, N. S., Johnson, C. E., Rogers, R. L., & Petrie, T. A. (1998). Social physique anxiety and disordered eating:
what’s the connection?. Addictive Behaviors, 23, 1–6.
Garner, D. M., & Olmstead, M. P. (1984). Manual for eating disorder inventory. Odessa, FL: Psychological Assess-
ment Resources.
E.V. Monsma, R.M. Malina / Psychology of Sport & Exercise 5 (2004) 447–460 458
Garner, D. M., Olmstead, M. P., & Polivy, J. (1983). Development and validation of a multidimensional eating dis-
order inventory for anorexia nervosa and bulimia. International Journal of Eating Disorders, 2, 15–34.
Gould, D., Jackson, S. A., & Finch, L. M. (1993). Sources of stress in national champion figure skaters. Journal of
Sport and Exercise Psychology, 15, 135–159.
Harris, R. (1986). On aspects of jumping. Skating, 63, 10–13.
Hausenblas, H. A., & Carron, A. V. (1999). Eating disorder indices and athletes: an integration. Journal of Sport and
Exercise Psychology, 21, 230–258.
Hausenblas, H. A., & Mack, D. E. (1999). Social physique anxiety and eating disorder correlates among female ath-
letic and non-athletic populations. Journal of Sport Behavior, 22, 502–513.
Himes, J. H., & Faricy, A. (2001). Validity and reliability of self-reported stature and weight of US adolescents.
American Journal of Human Biology, 13, 255–260.
Killen, J. D., Haywar, C., Litt, I., Hammer, L. D., Wilson, D. M., Taylor, C. B., Varaday, A., & Shisslack, C.
(1992). Is puberty a risk factor for eating disorders?. American Journal of Diseases in Children, 146, 323–325.
Leary, M. R. (1992). Self-presentational processes in exercise and sport. Journal of Sport and Exercise Psychology, 14,
339–351.
Leary, M. R., Tchividjian, L. R., & Kraxberger, B. E. (1994). Self presentation can be hazardous to your health:
impression management and health risk. Health Psychology, 13, 461–470.
Lerner, R. M. (1985). Adolescent maturational changes and psycho-social development: a dynamic interactional per-
spective. Journal of Youth and Adolescence, 14, 355–372.
Lohman, T. G., Roche, A. F., & Martonell, R. (Eds.) (1988), Anthropometric standardization reference manual.
Champaign, IL: Human Kinetics.
Malina, R. M. (1995). Anthropometry. In P. J. Maud, & C. Foster (Eds.), Physiological assessment of human fitness
(pp. 205–219). Champaign, IL: Human Kinetics.
Malina, R. M., Bouchard, C., & Bar Or, O. (2004). Growth, maturation, and physical activity. (2nd ed). Champaign,
IL: Human Kinetics.
Marsh, H. W. (1996). Construct validity of Physical Self Description Questionnaire responses: relations to external
criteria. Journal of Sport and Exercise Psychology, 18, 111–131.
Marsh, H. W. (1998). Age and gender effects in physical self-concepts for adolescent elite-athletes and non-athletes: a
multi-cohort–multi-occasion design. Journal of Sport & Exercise Psychology, 20(3), 237–259.
Marsh, H. W., Richards, G., Johnson, S., Roche, L. A., & Tremayne, P. (1994). Physical Self-Description Question-
naire: psychometric properties and multitrait-multimethod analysis of relations to existing instruments. Journal of
Sport and Exercise Psychology, 16, 270–305.
Martin, K. A., Rejeski, W. J., Leary, M. R., McAuley, E., & Bane, S. (1997). Is the Social Physique Anxiety Scale
really multidimensional? Conceptual and statistical arguments for a unidimensional model. Journal of Sport and
Exercise Psychology, 19, 357–367.
Monsma, E. V., & Malina, R. M. (in press). Anthropometry and somatotype of competitive female figure skaters 11–
22 years: Variation by competitive level and discipline. Journal of Sports Medicine and Physical Fitness.
Monsma, E. V., Malina, R. M., & Feltz, D. L. (2003). Puberty and Physical Self-Perceptions of Competitive Female
Figure Skaters. Manuscript submitted for publication.
Petrie, T. A. (1993). Disordered eating in female collegiate gymnasts: prevalence and personality/attitudinal corre-
lates. Journal of Sport and Exercise Psychology, 15, 424–436.
Petrie, T. A. (1996). Differences between male and female college lean sport athletes, non-lean sport athletes, and
nonathletes on behavioral and psychological indices on eating disorders. Journal of Applied Sport Psychology, 8,
218–230.
Reel, J. J., & Gill, D. L. (1996). Psychosocial factor related to eating disorders among high school and college female
cheerleaders. The Sport Psychologist, 10(180-20), 6.
Rezek, P. J., & Leary, M. R. (1991). Perceived control, drive for thinness and food consumption: anorexic tendencies
as displaced reactance. Journal of Personality, 59, 129–142.
Rucinski, A. (1989). Relationship of body image and dietary intake of competitive ice skaters. Journal of American
Dietary Association, 89, 98–100.
459 E.V. Monsma, R.M. Malina / Psychology of Sport & Exercise 5 (2004) 447–460
Scanlan, T. K., Stein, G. L., & Ravizza, K. (1991). An in-depth study of former elite figure skaters: III. Sources of
stress. Journal of Sport and Exercise Psychology, 13, 103–120.
Stoutjesdyk, D., & Jevne, R. (1993). Eating disorders among high performance athletes. Journal of Youth and Ado-
lescence, 22, 271–282.
Sundgot-Borgen, J. (1994). Risk and trigger factors for the development of eating disorders in female athletes. Medi-
cine and Science in Sports and Exercise, 26, 414–419.
Tabachnick, B., & Fidell, L. (1996). Using multivariate statistics. (3rd ed). New York: Harper Collins.
Taylor, G. M., & Ste.-Marie, D. M. (2001). Eating disorders symptoms in Canadian female pair and dance figure
skaters. International Journal of Sport Psychology, 32, 21–28.
Thompson, A. M., & Chad, K. E. (2002). The relationship of social physique anxiety to risk for developing an eating
disorder in young females. Journal of Adolescent Health, 31, 183–189.
Vadocz, E. A. (1999). The psychobiological profile of competitive female figure skaters. Unpublished doctoral disser-
tation, Michigan State University, East Lansing, USA.
Vadocz, E. A., Siegel, S. R., & Malina, R. M. (2002). Age at menarche in competitive figure skaters: variation by
competency and discipline. Journal of Sports Sciences, 20, 93–100.
Ziegler, P., Hensley, S., Roepke, J. B., Whitaker, S. H., Craig, B. W., & Drewnowski, A. (1998). Eating attitudes and
energy intakes of female skaters. Medicine and Science in Sports and Exercise, 30, 583–586.
Ziegler, P., Khoo, C. S., Sherr, B., Nelson, J. A., Larson, W. A., & Drewnowski, A. (1998). Body image and dieting
behaviors among elite figure skaters. International Journal of Eating Disorders, 4, 421–427.
E.V. Monsma, R.M. Malina / Psychology of Sport & Exercise 5 (2004) 447–460 460