European Eating Disorders Review

Eur. Eat. Disorders Rev. 14, 23–31 (2006)

An Application of Escape Theory

to Binge Eating
Sonja Blackburn1y , Lucy Johnston1*, Neville Blampied1,
Danielle Popp3 and Rachel Kallen2
1
University of Canterbury, Private Bag 4800, Christchurch, New Zealand
2
University of Connecticut, Connecticut, USA
3
Florida Atlantic University, Florida, USA

The application of Escape Theory (Heatherton & Baumeister, 1991)

to binge eating was studied in a non-clinical sample of 129 women.
Structural equation modelling (SEM) showed a good fit between
the Escape Model and the data. Perfectionism strongly predicted
aversive self-awareness which, in turn, predicted negative affect.
Negative affect predicted levels of avoidant coping which strongly
predicted levels of binge eating. Implications for understanding
and treating binge eating are considered. Copyright # 2006 John
Wiley & Sons, Ltd and Eating Disorders Association.

Keywords: binge eating; Escape Model

INTRODUCTION as high as 26% (Cooper & Fairburn, 1983), and 30%

in samples from weight control programmes
Binge eating is recognised as a significant problem
(Brody, Walsh, & Devlin, 1994; Spitzer et al., 1992).
within modern western societies. Although often
The importance of understanding the underlying
conceptualised as a symptom of clinical syndromes
causal processes of binge eating is well recognised,
such as bulimia nervosa and anorexia nervosa, binge
particularly with reference to how this guides treat-
eating has a unique definition (Diagnostic and Sta-
ment and prevention strategies.
tistical Manual of Mental Disorders, (DSM IV-TR);
Heatherton and Baumeister’s (1991) Escape Model
American Psychiatric Association (APA), 2000).
(see Figure 1) offered a promising contribution to
There is growing evidence that binge eating fre-
understanding the causal mechanisms underlying
quently occurs in the absence of other eating disor-
binge eating, emphasizing the function that binge
der diagnoses, and is perhaps best thought of as
eating serves for an individual. What is striking,
lying along a continuum of normal to disordered
however, is that while this model is well cited in
eating (Thelan, Mann, Pruitt, & Smith, 1987).
the literature, relatively few studies have directly
Estimates of prevalence rates for binge eating dis-
tested its applicability to binge eating. The present
order (BED) range from 2% to 5% (Bruce & Agras,
study specifically tested the applicability of Escape
1992; Fairburn, Hay, & Welch, 1993; Spitzer et al.,
theory in a non-clinical sample of women who binge
1993) but binge eating that occurs less frequently
eat, using structural equation modelling (SEM).
than the twice-weekly BED diagnostic requirement
SEM provided a means to distinguish sequential
is far more common, with estimates of prevalence
relations within the model and also to test the good-
ness of fit between the data and the hypothesized
* Correspondence to: Lucy Johnston, Associate Professor, model, through simultaneously testing various
Department of Psychology, University of Canterbury, Private pathways within the model. SEM is a largely a con-
Bag 4800, Christchurch, New Zealand. Tel: þ 64 3 364 2967. firmatory, rather than exploratory, technique. That
Fax: þ 64 3 364 2181. is, SEM is used to determine whether a certain model
E-mail: [email protected]
y
The reported research was conducted in partial fulfilment of is valid, rather than to find a suitable model. A struc-
the requirements for an MA in Psychology by the Sonja tural equation model implies a structure of the cov-
Blackburn. ariance matrix of the measures within the model.

Copyright # 2006 John Wiley & Sons, Ltd and Eating Disorders Association.
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/erv.675
24 S. Blackburn et al.

Once the model’s parameters have been estimated, In such a state, it is argued, meaningful interpreta-
the resulting model-implied covariance matrix can tions such as attributions, comparisons against stan-
then be compared to an empirical or data-based dards and implications of one’s actions are no longer
covariance matrix. If the two matrices are consistent readily accessible, and accordingly negative affect is
with one another, then the structural equation likely to be alleviated. That is, cognitive narrowing
model can be considered a plausible explanation leads to a reduction in, or escape from, the aversive
for relations between the measures. Accordingly, state of negative affect felt by an individual as a con-
we tested the theoretical model specified by sequence of heightened self-awareness.
Heatherton and Baumeister (1991) against the data This escape from negative affect may however
collected from a sample of women who completed precipitate a cascade of self-defeating behaviours,
measures pertaining to the constructs of the Escape such as binge eating. Efforts to escape from unplea-
Model. sant feelings through cognitive narrowing may
The Escape Model (Heatherton & Baumeister, erode the usual inhibitions around food (or any
1991) is built on concepts drawn from theories of other self-destructive behaviour), making the sus-
self-awareness and comparisons with one’s ideal ceptible person more willing to break their dietary
self, and has been applied to a number of self-defeat- rules, and may facilitate further escape through nar-
ing behaviours, including binge eating (Baumeister rowing attention to the actions and sensations
& Scher, 1988; Heatherton & Baumeister, 1991), sex- involved with eating. With additional cognitive nar-
ual masochism (Baumeister, 1990) and suicide rowing prior inhibitions around eating are eroded,
(Heatherton & Baumeister, 1991). Although not an and eating escalates into a full-blown binge episode.
inherently aversive state, self-awareness can be bur- A binge episode is, then, a result of the escape from
densome, particularly when an individual has high, negative affect through cognitive narrowing, rather
demanding, even perfectionist, standards and than a cause of that narrowing. Binge eating gener-
becomes aware that they are failing to meet their per- ally occurs within a context of aspirations to meet
sonal goals and ideals (Duval & Wicklund, 1972). If particular standards of weight or body image, and
these discrepancies are then attributed to internal deliberate efforts to reduce weight. Individuals for
aspects of the self, the individual is likely to experi- whom physical appearance, body size, shape and
ence negative affect, which is an aversive state. Indi- weight are important features of the self are likely
viduals are motivated to escape this negative to engage in dietary restraint of some form. Accord-
affective state. The Escape Model provides a frame- ingly, for these individuals, cognitive narrowing
work for viewing how individuals might escape the will reduce the usual restraints and inhibitions
aversive negative affective state. around eating, and increase susceptibility to binge
Each of the steps in the Escape Model is viewed as eating as a consequence of escape from negative
a choice point in a decision tree, with bingeing only affect through cognitive narrowing. Further, the
occurring if each step produces a particular outcome binge episode may provide the individual with an
(Heatherton & Baumeister, 1991). Perfectionist stan- opportunity to attribute their negative affect to the
dards per se are unlikely to lead to binge eating. If, binge episode rather than to the original (internal)
however, these perfectionist standards result in eva- source of distress and state of negative affect from
luative interpretations that highlight perceived which the individual was trying to escape (Polivy
shortcomings which are attributed to aspects of the & Herman, 1999), hence perpetuating the long-term
self (rather than to external factors), so creating a problem.
state of negative self-awareness, negative affect is Past research has furnished much information
likely to ensue. The individual is then motivated to about the differences between the binge-prone indi-
escape this negative affect. One way to reduce nega- vidual and the non-binger on those characteristics
tive affect is by directly reducing self-awareness identified in the Escape Model. Most research has
such that discrepancies between the self and focused on a single factor within the Escape Model,
relevant standards are no longer salient (Duval & although some researchers have shown binge-eaters
Wicklund, 1972). This reduction of self-awareness, to fit a profile of an escape-prone individual based in
or cognitive narrowing, is the type of escape consid- the Escape Model—that is, relative to non-bingers,
ered in the Escape Model. to be high on depression, anxiety, self-focus and per-
The central claim of the Escape Model is that self- fectionism, and to show a disposition to resorting to
awareness is reduced by cognitive narrowing, cognitive narrowing of attention but to be low in
through a focusing of attention on concrete aspects self-esteem (Beebe, Holmbeck, Albright, Noga, &
of the immediate environment (Baumeister, 1990). Decastro, 1995; Striegel-Moore, Silberstein, & Rodin,

Copyright # 2006 John Wiley & Sons, Ltd and Eating Disorders Association. Eur. Eat. Disorders Rev. 14, 23–31 (2006)
An Application of Escape Theory 25

1993; but see Tassava & Rudeman, 1999, for a con- of avoidant coping to predict bingeing was depen-
trary view). dent on levels of negative affect, consistent with
Most research has only considered individual the Escape Model. However, there has not been
components of the model. We briefly consider the any research that has tested the full Escape Model,
findings for each of the components of the and the relationships between each of the constructs,
model—perfectionism, negative self-awareness, with binge-eaters. This is the aim of the present
negative affect and cognitive narrowing. There is a research.
robust association between perfectionism and binge Our study provides a more rigorous test of the
eating (Fairburn et al., 1998; Fairburn, Cooper, Doll, Escape Model with respect to binge eating. We mea-
& Welch, 1999), although the influence of perfection- sured the various elements of the Escape Model
ism is mediated by other psychological factors amongst a non-clinical sample of self-defined female
including self-esteem, consistent with the Escape dieters and used SEM to test the fit of the data to the
Model (Pratt, Telch, Labouvie, Wilson, & Agras, Escape Model. We recruited participants who
2001; Vohs, Bardone, Joiner, & Abramson, 1999) defined themselves as dieters since the incidence
and anxiety (Davis, Claridge, & Fox, 2000). There is of binge eating is reported to be higher amongst
considerable evidence that binge eating is associated those attempting to control their weight (Brody
with lower levels of self-esteem (Eldredge, Wilson, et al., 1994; Spitzer et al., 1992). Our assessment
& Whaley, 1990; Gross & Rosen, 1988; Katzman & measures for the components of the Escape Model,
Wolchik, 1984; Mizes, 1988; Ross & Ivis, 1999), as described below, were selected after review of
although whether this lowered self-esteem is accom- the relevant past literature.
panied by a heightened sense of self-focus, as sug-
gested by the Escape Model, is less clear (Beebe
et al., 1995; Striegel-Moore et al., 1993; Tassava &
Ruderman, 1999). Negative affect is a salient predic- METHOD
tor of bulimic behaviours in adolescent girls (Stice &
Participants
Agras 1998), and of binge eating in overweight
women (Agras & Telch, 1998), although this rela- Women who identified themselves as dieters were
tionship is moderated by the tendency to use disen- recruited via advertisements on campus and in
gagement as a coping style (Henderson & Huon, community newspapers. As an incentive, partici-
2002). Not all forms of emotional distress trigger pants were entered into a prize draw. Of the 180
binge eating, however, only those which threaten questionnaires distributed, 145 were returned, giv-
self-esteem (Heatherton & Baumeister, 1991; ing a very high response rate of 81%. The data from
Heatherton et al., 1991), which is consistent with 16 women were however excluded as a conse-
the Escape Model. Supporting evidence for the cog- quence of errors or omissions in the completion
nitive narrowing or deconstruction process in binge- of the questionnaires. The final sample contained
eaters is indirect. Binge-eaters are characterised by a 77 students and 52 non-students (community sam-
greater level of dichotomous thinking than control ple). Although all the participants identified them-
subjects; they are more susceptible to salient external selves as dieters there was a wide range of scores
cues than internal cues, eat unusual and unpalatable on a measure of restrained eating —The Dutch Eat-
foods during a binge episode, and frequently report ing Behaviour Questionnaire (DEBQ-Res.; Van
feeling numb during a binge episode (Paxton & Strein et al., 1986). Scores ranged from 11 to 50 (pos-
Diggens, 1997), eating patterns that could be consid- sible range is from 10 to 50) and were normally dis-
ered to be indicative of a state of mind in which tributed about a mean of 33.69. The women had a
meaningful thought has been blocked. Individuals mean age of 29 years and 11 months (range 18–64
with eating disorders have also been shown to use years), with a higher mean for the community than
more avoidant coping strategies than controls (Ball the student sample (Ms ¼ 41.08 vs. 22.33 years;
& Lee, 2000). t(127) ¼ 11.66, p < 0.01). The mean Body Mass
While this literature does provide some evidence Index (BMI) for the sample was 25.54, but this
that binge-eaters have an escape-prone personality, was also significantly higher for the community
little attention has been paid to the processes than the student sample (Ms ¼ 28.89 vs. 23.25;
leading up to bingeing. Paxton and Diggens (1997) t(121) ¼ 6.46, p < 0.01). Overall the range of BMI
did consider the extent to which aversive self-aware- in our sample was from 17.53 to 41.40 and the mean
ness, negative affect and avoidance behaviours pre- BMI 25.54. Twelve participants (9%) has a BMI of
dicted binge eating. They concluded that the power less than 20, 89 participants (69%) a BMI of between

Copyright # 2006 John Wiley & Sons, Ltd and Eating Disorders Association. Eur. Eat. Disorders Rev. 14, 23–31 (2006)
26 S. Blackburn et al.

20 and 30, and 21 participants (17%) a BMI of energy may indeed be an artefact of disrupted eating
greater than 30. patterns, rather than a reflection of mood.
Cognitive Narrowing was measured using the avoi-
Materials dant coping subscales of the COPE (Carver, Scheier,
A battery of questionnaires was compiled to & Weintraub, 1989). The COPE is a multidimensional
measure the components of the Escape Model— inventory that can be used to assess both situational
perfectionism, aversive self-awareness, negative coping, and more general dispositional coping styles,
affect, cognitive narrowing and binge eating—as through 13 distinct subscales. The concept of escape,
detailed below.1 through cognitive narrowing, described within the
Perfectionism was measured using the Positive and Escape Model of binge eating refers to a motivated
Negative Perfectionism Scale (PANPS; Terry-Short, attempt to avoid the negative feelings associated with
Owens, Slade, & Dewey, 1995). This scale is a 40-item aversive self-awareness. Heatherton et al. (1991) con-
measure of perfectionism which comprises two sub- ceptualised this cognitive narrowing as an avoidant
types of perfectionist behaviour—that which arises style of coping, a tendency to cope with situations
as a function of positive reinforcement (positive per- by avoiding thinking about them. Typical strategies
fectionism), and that which arises as a function of used to facilitate this process include alcohol use, dis-
negative reinforcement (negative perfectionism). traction, mental and behavioural disengagement or
Aversive self-awareness was measured using two indeed eating. Typically, coping strategies such as
subscales of the self-consciousness scale-revised these focus on regulating the emotional response to
(SCSR; Fenigstein, Scheier, & Buss, 1975). This scale stress, as opposed to focusing attention on the actual
incorporates three subscales, public self-conscious- source of the stress.
ness (awareness of the reaction of others to the self), In the present study, the dispositional version of
private self-consciousness (attention to one’s inner the COPE was administered, and scores on the sub-
thoughts and feelings) and social anxiety (discom- scales measuring denial, mental disengagement,
fort in the presence of others), though only the public behavioural disengagement, and use of alcohol
and private self-consciousness sub-scales were used and drugs summed to provide a measure of avoi-
as our measure of aversive self-awareness, consis- dant coping. Denial refers to direct attempts made
tent with past research (Eichstaedt & Silvia, 2003). to deny the reality of the stressful event. Mental dis-
Negative Affect was measured using the Hospital engagement refers to psychological disengagement
Anxiety and Depression Questionnaire (HADS; from the goal that the stressor is interfering with
Zigmond & Snaith, 1983), a 14-item scale designed through methods such as daydreaming, sleeping
as a brief measure of the presence and severity of or self-distraction. Behavioural disengagement
both anxiety and depression symptoms. As its per- refers to a tendency to give up or withdraw effort
formance as a screening device is enhanced when from any attempts to attain the goal that the stressor
scores on the two subscales are summed (Ravazi, is interfering with.
Delvaux, Farvacques, & Robaye, 1990), we used a Bingeing was assessed using the binge scale of the
total score across the two subscales as our measure Bulimia Test (BULIT; Smith & Thelen, 1984). The
of negative affect. The HADS differs from other mea- BULIT was constructed by comparing responses of
sures of mood state primarily in its reliance on anhe- both bulimic subjects and female college students
donia as a symptom, rather than somatic symptoms against clinical judgments based on diagnostic inter-
such as loss of energy or tiredness. This shift in views. Scale scores were found to be a reliable pre-
emphasis is based on the idea that symptoms such dictor of individual bulimic symptoms when
as low energy, fatigue and sleep disturbance may compared against independent clinical interviews.
be attributable to physical illness rather than to The Binge subscale contains 14 items pertaining to
mood state per se (Zigmond & Snaith, 1983). This actual bingeing behaviour, such as the amount of
rationale makes the HADS an appropriate measure food consumed, speed of eating and eating to point
for populations with restricted or disordered eating of physical discomfort as well as items relating to
patterns, as symptoms such as fatigue and flagging loss of control over eating.

1
Participants also completed the DEBQ as detailed above and Procedure
the Rosenberg Self-Esteem Scale (RSE; Rosenberg, 1965). As
the RSE was not a measure of any of the components of the Participants were mailed a test booklet comprising
Escape Model, however, it is not reported here. Full details of the questionnaires detailed above, a demographic
this measure can be obtained from the corresponding author. data sheet, and an information sheet and consent

Copyright # 2006 John Wiley & Sons, Ltd and Eating Disorders Association. Eur. Eat. Disorders Rev. 14, 23–31 (2006)
An Application of Escape Theory 27

Table 1. Means and standard deviations for all variables

Total Sample N ¼ 129 Community subset N ¼ 52 Student subset N ¼ 77

Mean SD Mean SD Mean SD

Perfectionism 134.63 20.38 133.71 22.18 135.25 19.20
Aversive self-awareness 30.91 7.98 28.75 8.10 32.38 7.62
Negative affect 13.74 6.19 14.21 6.90 13.43 5.68
Avoidant coping 27.87 6.18 27.25 5.53 28.29 6.58
Binge eating 32.73 10.91 35.96 12.21 30.55 9.41

form, and a prepaid reply envelope. To control for (t(127) ¼ 2.84, p < 0.05; Ms ¼ 35.96 vs. 30.55), indicat-
practice or priming effects, different versions of the ing that the community sample reported more
booklet were developed by counter-balancing the bingeing-related behaviours than did the student
order of the questionnaires. group.
To distinguish between those women who binged
occasionally and those who represented the upper
spectrum of bingeing behaviour, a cut-off score
RESULTS
derived from the normative mean plus one standard
Means and standard deviations for each of the mea- deviation (Stein & Brinza, 1989) classified women
sured variables are presented in Table 1. scoring above 34 as high binge-eaters and those scor-
Scores on all of the reported measures were nor- ing 34 or below as low binge-eaters. The majority of
mally distributed. Comparisons (t-tests) of the stu- women in our sample were, accordingly, occasional
dent and community samples on each measure bingers, but some 51 (39.5%) did score over 34, indi-
revealed only two significant differences. Given cative of being within the upper spectrum of binge-
the relatively few differences between these two ing behaviour. Further, it should be noted that
sample groups, however, they were combined for within the student population 31% scored above 34
the SEM analysis. The student sample scored higher on the BULIT, but within the community sample
on the self-consciousness scale (SCSR) than did the some 52% scored above 34.
community sample (t(127) ¼ 2.06, p < 0.05; Ms ¼ Using the analysis of moment structures (AMOS)
41.95 vs. 38.37), indicating that the student group software, we tested the theoretical predictions of
had a higher degree of self-consciousness on average Heatherton and Baumeister’s Escape Model. As we
than the community group. The community sample were testing a predetermined theoretical model
scored higher on the BULIT measure of binge eating (see Figure 1), we did not make any re-specifications

Perfectionism Aversive Self- Negative Cognitive Binge

Awareness Affect Narrowing Eating

Figure 1. The Escape Model, applied to binge eating (after Heatherton & Baumeister, 1991)

e1 e2
1 1

Public Self-Cons Private Self-Cons

Perfectionism Aversive Self Negative Affect Cognitive Narrowing Binge Eating

Awareness
1 1 1 1

U V W X

Figure 2. Structural Model

Copyright # 2006 John Wiley & Sons, Ltd and Eating Disorders Association. Eur. Eat. Disorders Rev. 14, 23–31 (2006)
28 S. Blackburn et al.

Table 2. Covariances between model variables

PANPS HADS BULIT COPE PRIVATE PUBLIC

PANPS 412.249
HADS 59.742 37.973
BULIT 54.651 27.977 118.089
COPE 42.656 16.594 10.143 37.866
PRIVATE 38.586 10.122 1.971 5.105 23.775
PUBLIC 45.142 6.740 5.572 4.218 11.061 17.359

to the model. The structural model tested is shown in efficient ¼ 0.91). In turn aversive self-awareness
Figure 2. The input to the analysis was the observed was predictive of negative affect (standardised co-
data from our sample and there was no missing data. efficient ¼ 0.52) and negative affect of cognitive nar-
The covariance matrix is shown in Table 2. There rowing (standardised co-efficient ¼ 0.83). Cognitive
were five observed, endogenous variables—nega- narrowing was very strongly predictive of binge eat-
tive affect (measured by the HADS), cognitive nar- ing (standardised co-efficient ¼ 0.95). In addition,
rowing (measured by the COPE), binge eating there was also a negative path coefficient between
(measured by the BULIT), and public and private binge eating and cognitive narrowing (standardised
self-consciousness (measured by subscales of the co-efficient ¼
0.94).
SCSR)—and one observed exogenous variable—
perfectionism (measured by the PANPS). There
was a single unobserved endogenous variable—
DISCUSSION
aversive self-awareness—that was composed of
indicators representing the measured variables of The present research tested the applicability of the
public and private self-consciousness. Higher Escape Model (Heatherton & Baumeister, 1991) to
aversive self-awareness predicted higher public binge eating in a non-clinical sample of women.
(standardized co-efficient ¼ 0.55) and private (stan- Our participants were assessed on each of the factors
dardized co-efficient ¼ 0.44) self-consciousness, and in the Escape Model and then SEM was used to esti-
public and private self-consciousness were corre- mated the goodness of fit of the model to the
lated at r(129) ¼ 0.544, p < 0.001. observed data. Our analyses demonstrated that the
Support was found for the hypothesised model in Escape Model was indeed a good fit to the data, sug-
terms of the
2 test statistic,
2(7) ¼ 12.18, p < 0.095, gesting that the model does provide a framework for
the comparative fit index (CFI), CFI ¼ 0.97, and the understanding the causal processes leading to binge
root mean square error of approximation (RMSEA); eating and explaining the function binge eating
RMSEA ¼ 0.076, PCLOSE ¼ 0.235. serves for an individual.
The standardised regression coefficients are The individual path coefficients offered support
shown in Figure 3. All of the path coefficients for the relationships between the model variables
between model variables were highly significant predicted in the Escape model—stronger perfec-
(p < 0.001). Perfectionism was strongly predictive tionist ideals lead to higher levels of aversive self-
of aversive self-awareness (standardised co- awareness which, in turn, lead to higher negative

.39

e1 e2

Public Self-Cons Private Self-Cons

.58 .44

.95
.91 Aversive Self .52 .83
Perfectionism Negative Affect Cognitive Narrowing -.94 Binge Eating
Awareness

U V W X

Figure 3. Structural Model: standardised estimate

Copyright # 2006 John Wiley & Sons, Ltd and Eating Disorders Association. Eur. Eat. Disorders Rev. 14, 23–31 (2006)
An Application of Escape Theory 29

affect. Higher negative affect lead to greater cogni- measure of the extent to which this self-awareness is
tive narrowing, which is consistent with the indivi- unpleasant for the individual may create a stronger
dual attempting to escape from feelings of negative measure of aversive self-awareness than is available
affect. Cognitive narrowing then predicted levels of in the present study. The relative weighting of pri-
binge eating. It is important to also note that there vate and public self-evaluation might also be consid-
was support for the bi-directional nature of the rela- ered within the population of binge-eaters. In the
tionship between cognitive narrowing and binge present research, the contributions of private and
eating. However, the relationship between binge public self-consciousness to the measure of negative
eating and cognitive narrowing was a negative self-awareness were equally weighted, but it is pos-
one. Higher levels of binge eating predicted lower sible that in the context of binge eating private self-
levels of cognitive narrowing. Rather than binge eat- consciousness may be of greater importance (since
ing and cognitive narrowing operating in negative many binge episodes are conducted in private) and
feedback cycle, our data suggest that binge eating so the contribution of private self-consciousness
reduces cognitive narrowing. This may, however, should be greater than that of public self-conscious-
lead to enhanced negative self-awareness and a ness in the measurement of negative self-awareness.
perpetuation of the negative affect-cognitive nar- In the development of the Escape Model
rowing-binge eating cycle. The nature of this rela- (Heatherton & Baumeister, 1991), it was assumed
tionship warrants further research. One issue that dietary restraint was an important precursor
worthy of further investigating is the role of purging to binge eating. Accordingly, our sample was
behaviour. It is possible that purging after binge- recruited from women who self-identified as diet-
ing may result in an individual feeling ‘empty and ers. However, our participants’ scores on a measure
clean’ and a reduction of negative affect, leading to of restrained eating—The Dutch Eating Behaviour
a reduction in cognitive narrowing and binge eating. Questionnaire (DEBQ-Res.; Van Strein et al.,
Alternatively, however, purging may increase aver- 1986)—covered the entire range of possible scores
sive self-awareness in a manner similar to binge eat- and were normally distributed. Further, there was
ing itself and so perpetuate the negative cycle no significant correlation between level of dietary
predicted by the Escape Model. Given the close links restraint and level of binge-eating (r(129) ¼ 0.073).
between bingeing and purging it is important for This seeming independence of dieting and bingeing
future research to consider the incidence, and impact challenges the assumption of the Escape Model that
of, purging as well as binge-eating behaviour. dietary restraint is an important precursor to binge
Overall the goodness of fit of the Escape Model to eating and suggests that the Escape Model can pre-
our data was high and each of the path coefficients dict binge eating behaviours regardless of level of
between model variables was highly significant. dietary restraint. Whilst the Escape Model may
The path between aversive self-awareness and nega- apply to binge eating irrespective of dieting status,
tive affect was, however, relatively weak. Aversive co-existing dietary restraint may still serve to exacer-
self-awareness was the only exogenous variable in bate the conditions that lead to emotional distress
our model. That is, it was the only measure that hence creating a wider range of conditions promot-
was derived indirectly through more than one mea- ing escape and more powerful motivation for
surement of other constructs (public and private escape. Whether a self definition of being a dieter,
self-consciousness) rather than derived directly regardless of actual levels of dietary restraint, as in
from a single measure. We were guided in our choice the present sample, is an important factor is unclear
of measurement tools for this variable, and indeed and further research is needed in which the Escape
for all our variables, by past literature, but future Model for binge eating is tested amongst a wider
research may consider whether aversive self-aware- sample, including both those who do not self-define
ness might be better captured by an alternative mea- as dieters and also male participants. Indeed, it
sure, or measures, which would, in turn, increase the should be noted that even within female dieters
path coefficient to negative affect. The measures our sample was largely restricted to those of normal
used in the present study—public and private self- weight (as indexed by BMI) and further considera-
consciousness—may focus too strongly on the self- tion needs to be given to dieters who are of high or
awareness component of negative self-awareness, low weight as this may give rise to higher levels of
whether that be on awareness of one’s own feelings negative affect.
(private self-consciousness) or the reactions of It is noteworthy that our participants were a
others (public self-consciousness), rather than on community-based sample, either drawn from a
the evaluation of that awareness. Inclusion of some student population or a wider community-based

Copyright # 2006 John Wiley & Sons, Ltd and Eating Disorders Association. Eur. Eat. Disorders Rev. 14, 23–31 (2006)
30 S. Blackburn et al.

population. The only criterion for inclusion in our REFERENCES

sample was a self-definition as a ‘dieter,’ although
as seen from scores on the DEBQ-Res the women Agras, W., & Telch, C. (1998). The effects of caloric
in our sample showed a wide range of dieting levels. deprivation and negative affect on binge eating in
The incidence of upper-level binge eating (Stein & obese binger-eating disordered women. Behaviour
Therapy, 29, 491–503.
Brinza, 1989) in our sample (35%) is consistent with American Psychiatric Association. (2000). Diagnostic
that seen in past research (Brody et al., 1994; Cooper and statistical manual of mental disorders (4th ed.).
& Fairburn, 1983; Spitzer et al., 1992). It is perhaps a Washington, DC: American Psychiatric Association.
little surprising that higher levels of bingeing were Ball, K., & Lee, C. (2000). Relationships between psycho-
seen among the community than the student popu- logical stress, coping and disordered eating: A review.
lations in our sample but given the voluntary nature Psychology and Health, 14, 1007–1035.
Baumeister, R. F. (1990). Suicide as escape from self.
of participation in the research and the self-defini-
Psychological Review, 97, 90–113.
tion of ‘dieter’ required for participation, we need Baumeister, R. F., & Scher, S. J. (1988). Self-defeating
to be cautious in over-interpreting this finding, but behaviour patterns among normal individuals:
it is one that calls for future investigation. Review and analysis of common self-destructive
In summary, this paper presents the first test of tendencies. Psychological Bulletin, 104, 3–22.
the Escape Model applied to binge eating using Beebe, D. W., Holmbeck, G. N., Albright, J. S., Noga, K., &
SEM. The high goodness of fit between the theore- Decastro, B. (1995). Identification of ‘Binge-Prone’
women: An experimentally and psychometrically
tical model and our data offers strong support for validated cluster analysis in a college population.
the use of the Escape Model in both explaining and Addictive Behaviours, 20, 451–462.
predicting binge eating behaviour, and in the Brody, M. J., Walsh, B. T., & Devlin, M. J. (1994). Binge
developments of interventions to reduce binge eat- eating disorder: Reliability and validity of a new
ing behaviour. This has implications for the clinical diagnostic category. Journal of Consulting and Clinical
management of binge eating, suggesting that ther- Psychology, 62, 381–386.
Bruce, B., & Agras, W. S. (1992). Binge-eating in females:
apy must focus on alternative strategies for mana- A population based investigation. International Journal
ging negative affect as much as on reducing dieting of Eating Disorders, 12, 365–373.
and controlling urges to binge. Current treatments Carver, C. S., Scheier, M. F., & Weintraub, J. K. (1989).
focus on addressing erratic eating patterns and dis- Assessing coping strategies: A theoretically-based
torted cognitions surrounding weight and shape. approach. Journal of Personality & Social Psychology,
Equally important, however, is the acknowledge- 56, 267–283.
Cooper, P. J., & Fairburn, C. G. (1983). Binge-eating and
ment that a given behaviour, no matter how see-
self-induced vomiting in the community: A prelimin-
mingly maladaptive, may provide an important ary study. British Journal of Psychiatry, 142, 139–144.
function for the individual. Escape theory suggests Davis, C., Claridge, G., & Fox, J. (2000). Not just a pretty
that binge eating does indeed serve such a func- face: Physical attractiveness and perfectionism in the
tion, by providing a means of escape from distres- risk for eating disorders. International Journal of Eating
sing emotions that arise in the context of negative Disorders, 27, 67–73.
self-evaluation. It would seem reasonable, then, Duval, S., & Wicklund, R. A. (1972). A theory of objective
self-awareness. New York: Academic Press.
to assume that the cycle of negative self-evaluation Eichstaedt, J., & Silva, P. J. (2003). Noticing the
and negative affect should be as much a focus self: Implicit assessment of self-focus attention using
of treatment as the bingeing behaviour itself. If word recognition latencies. Social Cognition, 21,
these underlying causal mechanisms are left untre- 349–361.
ated then there is a risk that the individual will Eldredge, K., Wilson, G. T., & Whaley, A. (1990). Failure,
relapse or engage in potentially more destructive self-evaluation and feeling fat in women. International
Journal of Eating Disorders, 9, 179–190.
escapist behaviours. The reality of this risk has Fairburn, C. G., Cooper, Z., Doll, H. A., & Welch, S. L.
been borne out by research showing that over a (1999). Risk factors for anorexia nervosa: Three
period of time, reductions in disordered eating pat- integrated case control comparisons. Archives of Gen-
terns were accompanied by increases in alcohol eral Psychiatry, 56, 468–476.
and drug abuse in bulimics and binge-eaters Fairburn, C. G., Doll, H. A., Welch, S. L., Hay, P. J.,
(Yager, Landsverk, Edelstein, & Jarvik, 1988). Davies, B. A., & O’Connor, M. E. (1998). Risk factors
for binge eating disorder: A community based, case
Long-term outcome may therefore be improved if
controlled study. Archives of General Psychiatry, 55,
current treatments were to incorporate strategies 425–432.
that interrupt the ruminative self-comparison pro- Fairburn, C. G., Hay, P. J., & Welch, S. L. (1993). Binge
cess and offer more adaptive methods of dealing eating and bulimia nervosa: Distribution and determi-
with emotional distress. nants. In C. G. Fairburn, & G. T. Wilson (Eds.), Binge

Copyright # 2006 John Wiley & Sons, Ltd and Eating Disorders Association. Eur. Eat. Disorders Rev. 14, 23–31 (2006)
An Application of Escape Theory 31

eating: Nature, assessment and treatment (pp. 123–143). Spitzer, R. L., Devlin, M., Walsh, B. T., Hasin, D., Wing, R.,
New York: Guilford Press. Marcus, M., Stunkard, A., Wadden, T., Yanovski, S.,
Fenigstein, A., Scheier, M. F. & Buss, A. H. (1975). Public Agras, S., Michell, J., & Nonas, C. (1992). Binge eating
and private self-consciousness: Assessment and theory. disorder: A multisite field trial of the diagnostic
Journal of Consulting and Clinical Psychology, 43, 522–527. criteria. International Journal of Eating Disorders, 11,
Gross, J., & Rosen, J. C. (1988). Bulimia in adolescents: 191–203.
Prevalence and psychosocial correlates. International Spitzer, R. L., Yanovski, S., Wadden, T., Wing, R., Marcus,
Journal of Eating Disorders, 7, 51–61. M. D., Stunkard, A., Devlin, M., Mitchell, J., Hasin, D.,
Heatherton, T. F., & Baumeister, R. F. (1991). Binge eating & Horne, R. L. (1993). Binge eating disorder: Its further
as an escape from self-awareness. Psychological Bulle- validation in a multisite study. International Journal of
tin, 110, 86–108. Eating Disorders, 13, 137–153.
Heatherton, T. F., Herman, C. P., & Polivy, J. (1991). Stein, D. M., & Brinza, S. R. (1989). The bulimia test: Factor
Effects of physical threat and ego threat on eating structure in junior high school, high school, and
behaviour. Journal of Personality and Social Psychology, college women. International Journal of Eating Disorders,
60, 138–143. 8, 225–230.
Henderson, N. J., & Huon, G. F. (2002). Negative affect Stice, E., & Agras, W. S. (1998). Predicting onset and
and binge eating in overweight women. British Journal cessation of bulimic behaviours during adolescence: A
of Health Psychology, 7, 77–87. longitudinal grouping analysis. Behaviour Therapy, 29,
Katzman, M. A., & Wolchik, S. A. (1984). Bulimia and 257–276.
binge eating in college women: A comparison of Striegel-Moore, R. H., Silberstein, L. R., & Rodin, J.
personality and behavioural characteristics. Journal of (1993). The social self in bulimia nervosa: Public
Consulting and Clinical Psychology, 52, 423–428. self-consciousness, social anxiety and perceived
Mizes, J. S. (1988). Personality characteristics of bulimic fraudulence. Journal of Abnormal Psychology, 102, 297–
and non-eating disordered female controls: A cogni- 303.
tive behavioural perspective. International Journal of Tassava, S. H., & Ruderman, A. J. (1999). Application of
Eating Disorders, 7, 541–550. escape theory to binge eating and suicidality in college
Paxton, S. J., & Diggens, J. (1997). Avoidance coping, women. Journal of Social and Clinical Psychology, 18,
binge eating and depression: An examination of the 450–466.
escape theory of binge eating. International Journal of Terry-Short, L. A., Owens, R. G., Slade, P. D., & Dewey,
Eating Disorders, 22, 83–87. M. E. (1995). Positive and negative perfectionism.
Polivy, J., & Herman, C. P. (1999). The effects of Personality and Individual Differences, 18, 663–668.
resolving to diet on restrained and unrestrained Thelen, M. H., Mann, L. L., Pruitt, J., & Smith, M. (1987).
eaters: The ‘false hope syndrome’. International Journal Bulimia: Prevalence and component factors in
of Eating Disorders, 26, 434–447. college women. Journal of Psychosomatic Research, 31,
Pratt, E. M., Telch, C. F., Labouvie, E. W., Wilson, G. T., & 73–78.
Agras, W. S. (2001). Perfectionism in women with Van Strein, T., Frijters, J. E. R., Bergers, G. P. A., & Defares,
binge eating disorder. International Journal of Eating P. B. (1986). The Dutch eating behaviour questionnaire
Disorders, 29, 177–186. (DEBQ) for assessment of restrained, emotional and
Razavi, D., Delvaux, N., Farvacques, C., & Robaye, E. external eating behaviour. International Journal of
(1990). Screening for adjustment disorders and major Eating Disorders, 5, 295–315.
depressive disorders in cancer patients. British Journal Vohs, K. D., Bardone, A. M., Joiner, T. E., & Abramson,
of Psychiatry, 156, 79–83. L. Y. (1999). Perfectionism, perceived weight status
Rosenberg, M. (1965). Society and the adolescent self-image. and self-esteem interact to predict bulimic symptoms:
Princeton, NJ: Princeton University Press. A model of bulimic symptom development. Journal of
Ross, H. E., & Ivis, F. (1999). Binge eating and substance Abnormal Psychology, 108, 695–700.
use among male and female adolescents. International Yager, J., Landsverk, J., Edelstein, C. K., Jarvik, M. E.
Journal of Eating Disorders, 26, 245–260. (1988). A 20-month follow-up study of 628 women
Ruderman, A. J., & Grace, P. S. (1988). Bulimics and with eating disorders: II. Course of associated symp-
restrained eaters: A personality comparison. Addictive toms and related clinical features. International Journal
Behaviours, 13, 359–368. of Eating Disorders, 7, 503–513.
Smith, M. C., & Thelen, M. H. (1984). Development and Zigmond, A. S., & Snaith, R. P. (1983). The hospital
validation of a test for bulimia. Journal of Consulting anxiety and depression scale. Acta Psychiatrica Scandi-
and Clinical Psychology, 52, 863–872. navia, 67, 361–370.

Copyright # 2006 John Wiley & Sons, Ltd and Eating Disorders Association. Eur. Eat. Disorders Rev. 14, 23–31 (2006)