Pergamon
01N)5-7967(94)E0031 -D
Behav. Res. Ther. Vol. 33, No. I, pp. 85-93, 1995
Copyright © 1994 ElsevierScienceLtd
Printed in Great Britain. All rights reserved
0005-7967/95 $7.00 + 0.00
BODY IMAGE, BODY DYSPHORIA, AND DIETARY
RESTRAINT: FACTOR STRUCTURE
IN NONCLINICAL SUBJECTS
DONALD A. WILLIAMSON l, SUSAN E. BARKER I , LISA J. BERTMAN I
and DAVID H. GLEAVES2
~Department of Psychology, Louisiana State University, Baton Rouge, LA 70803-5501 and
2The Renfrew Foundation, Philadelphia, Pa, U.S.A.
(Received 31 March 1993: received for publication 20 April 1994)
Summary--The principal aim of this study was to examine the factor structure of several assessment
methods used to measure dietary restraint, body dissatisfaction, and body image. Factor analysis was
employed to identify and confirm the primary constructs measured by these assessment methods. A total
of 206 undergraduate women were recruited as subjects. This sample was divided into two subsets of 100
and 106 subjects. On the first subset, principle components analysis identified three factors: body
dysphoria, dietary restraint, and body image. With the second subset of subjects, confirmatory factor
analysis cross-validated this factor structure. A two factor solution, body dysphoria and dietary restraint,
was identified and confirmed when the body image measure was converted to a self-minus-ideal
discrepancy score. These findings are discussed in relation to the definition of control groups to be used
in studies of anorexia and bulimia nervosa. Guidelines for the selection of measures for each of the three
factors also are presented.
INTRODUCTION
Recent theories of anorexia and bulimia nervosa have emphasized three causal variables: body
dysphoria, body image disturbances, and dietary restraint (Attie& Brooks-Gunn, 1989; Rosen,
1992; Thompson, 1992; Williamson, 1990). Investigations of samples of nonclinical young women
have found a "normative discontent" with body size and high rates of dieting (Cash, Winstead &
Janda, 1986; Rodin, Silberstein & Striegel-Moore, 1985). Many young women express concerns
about dieting and body size/appearance which are comparable to the level of concern expressed
by women diagnosed with anorexia and bulimia nervosa (Polivy & Herman, 1985; Silberstein,
Striegel-Moore & Rodin, 1987; Thompson, 1992).
Two recent studies of the psychopathology of bulimia nervosa have supported the idea that
bulimia is a complex, multifactorial syndrome (Gleaves, Williamson & Barker, 1993; Tobin,
Johnson, Steinberg, Staats & Dennis, 1991). Both of these factor analytic studies found three
independent factors which described the syndrome of bulimia nervosa: (a) affective and personality
disturbances, (b) bulimic behaviors, and (c) dietary restraint. Gleaves et al. (1993) identified a
fourth factor, body dissatisfaction. This research suggests that it may be possible to isolate
psychopathological features of anorexia and bulimia nervosa that are unique to these syndromes
and those that are unique to each disorder.
In recent years, an increasing number of studies have attempted to identify such factors by the
inclusion of control groups without an eating disorder, but which have elevated scores on measures
of dietary restraint (e.g. Counts & Adams, 1985; Lindholm & Wilson, 1988; Laessle, Tuschl, Waadt
& Pirke, 1989) or weight preoccupation (Garner, Olmsted, Polivy & Garfinkel, 1984). Nonclinical
samples of women with a disturbed body image have also been used in treatment studies (e.g.
Rosen, Saltzberg & Srebnik, 1989). These studies have typically included an eating disorder group,
a symptomatic group (e.g. high dietary restraint), and a non-eating disorder/non-symptomatic
group. Dependent variables have ranged from measurement of body size estimation (Lindholm &
Wilson, 1988) to a level of depression (Laessle et al., 1989).
There is merit in the use of this research paradigm. Using this basic strategy may make it possible
to 'map out' the probable determining variables of the many psychopathological features of
BRT 33/1~O 85