ArticlePDF Available

Motivations to eat: Scale development and validation

Authors:

Abstract and Figures

Objective: To validate a measure of psychological motivations to eat based on a four-category model of motivations for alcohol use (Cooper, 1994). Motivations specified by this model are: to cope with negative affect, to be social, to comply with others’ expectations, and to enhance pleasure.Method: In Study 1, 40 respondents were queried in an open-ended format about their reasons for eating; responses were content-coded to determine if they fit into the four theorized categories. In Study 2, an item pool was generated based on responses from Study 1, and random halves of a sample of 812 college students were used to test and then validate the hypothesized factor structure.Results: As expected, the final inventory yielded the four theorized categories. The factor structure was generally invariant across gender, and the resulting Motivations to Eat subscales uniquely predicted restrictive eating, bingeing, and purging.Discussion: Prior eating research has focused mainly on coping and compliance motivations. The present study identified four distinct motivations to eat that potentially are important for understanding healthy and disordered eating.
Content may be subject to copyright.
Motivations to eat: Scale development
and validation
Benita Jackson,
a,*
M. Lynne Cooper,
b
Laurie Mintz,
c
and Austin Albino
d
a
Channing Laboratory, Brigham & WomenÕs Hospital/Harvard Medical School;
Harvard School of Public Health, 181 Longwood Avenue, Boston, MA, USA
b
Department of Psychology, 105 McAlester Hall, University of Missouri-Columbia,
Columbia, MO 65211, USA
c
Department of Educational and Counseling Psychology, 16 Hill Hall, University of Missouri-Columbia,
Columbia, MO 65211-2130, USA
d
Department of Psychological Sciences, 200 S. 7th St., Rm 124, University of Missouri-Columbia,
Columbia, MO 65211, USA
Abstract
Objective: To validate a measure of psychological motivations to eat based on a four-cat-
egory model of motivations for alcohol use (Cooper, 1994). Motivations specified by this mod-
el are: to cope with negative affect, to be social, to comply with othersÕexpectations, and to
enhance pleasure.
Method: In Study 1, 40 respondents were queried in an open-ended format about their rea-
sons for eating; responses were content-coded to determine if they fit into the four theorized
categories. In Study 2, an item pool was generated based on responses from Study 1, and ran-
dom halves of a sample of 812 college students were used to test and then validate the hypoth-
esized factor structure.
Results: As expected, the final inventory yielded the four theorized categories. The factor
structure was generally invariant across gender, and the resulting Motivations to Eat subscales
uniquely predicted restrictive eating, bingeing, and purging.
Discussion: Prior eating research has focused mainly on coping and compliance motiva-
tions. The present study identified four distinct motivations to eat that potentially are impor-
tant for understanding healthy and disordered eating.
Ó2003 Elsevier Science (USA). All rights reserved.
Journal of Research in Personality 37 (2003) 297–318
www.elsevier.com/locate/jrp
JOURNAL OF
RESEARCH IN
PERSONALITY
*
Corresponding author. Fax: +617-525-2578.
E-mail address: benita.jackson@channing.harvard.edu (B. Jackson).
0092-6566/03/$ - see front matter Ó2003 Elsevier Science (USA). All rights reserved.
doi:10.1016/S0092-6566(02)00574-3
1. Introduction
Eating is a physiological requirement, insofar as it provides nutrients for proper
growth and functioning of the major systems of the human body (e.g., cardiovas-
cular, immune, and pulmonary). Even beyond physiological requirements for suste-
nance, psychological motivations play an important role in the initiation of both
healthy and disordered eating. For example, eating can be motivated by both po-
sitive (e.g., happiness) and negative (e.g., anger, depression) emotions (Arnow,
Kenardy, & Agras, 1995; Heatherton & Baumeister, 1991; Macht & Simons,
2000; Sherwood, Crowther, Wills, & Ben-Porath, 2000), as well as by a desire to
nurture oneself (Leham & Rodin, 1989). Eating can also be motivated by internal-
ized social cues, norms, and expectations about food consumption (Leary, Tchividj-
ian, & Kraxberger, 1994; Roth, Herman, Polivy, & Pliner, 2001) or by cultural
practices that strengthen feelings of social connection, such as celebrations and
other social functions involving food (Pliner & Chaiken, 1990). Such motivations
to initiate eating have been the focus of some discussion and empirical research,
but these motivations most often have been discussed in isolation from one
another.
Allison (1994) provided an exhaustive review of existing inventories pertaining to
eating behaviors and eating disorders; none of the over 100 inventories reviewed was
designed to assess a range of psychological motivations to eat. Nevertheless, a few
inventories examined one or two aspects of psychological motivations concerning
food-related behaviors. For example, the Emotional Eating Scale (Arnow et al.,
1995) assesses an important psychological motivation for eating: as a response to
negative emotions. Two of the three scales in the Dutch Eating Behavior Question-
naire (van Strien, Frijters, Bergers, & Defares, 1986) tap motivations to eat: Emo-
tional Eating (eating in response to negative emotions) and External Eating
(eating in response to external sensory cues, such as the smell and appearance of
food). Finally, the Eating Inventory (Stunkard & Messick, 1985; cf. Stunkard &
Messick, 1988) has three scales, including one scale (disinhibition of control) that in-
cludes some items reflecting psychological reasons to eat. However, the disinhibition
of control scale includes items that reflect psychological motivations to eat (e.g.,
‘‘When I feel lonely, I console myself by eating’’) along with other non-motivation
items (e.g., ‘‘My weight has hardly changed at all in the last ten years’’). Thus the
scale does not provide a pure measure of motivations, nor was it intended to. Since
AllisonÕs 1994 book, Steptoe, Pollard, and Wardle published a measure of motiva-
tions that are associated with eating behavior (1995). However, their measure is de-
signed to assess what motivates food choices, once a person already has decided to
eat in the first place. In summary, though there are scales related to psychological
motivations to eat, none are theoretically derived for measuring exclusively a range
of motivations to initiate eating.
The relative lack of systematic theoretical and empirical attention to psychologi-
cal motivations in the field of eating disorders is surprising, especially when one ex-
amines the literature pertaining to psychological motivations for alcohol use. This
literature is particularly relevant considering the relation between eating disorders
298 B. Jackson et al. / Journal of Research in Personality 37 (2003) 297–318
and alcohol use. Specifically, both alcohol use and eating disorders can be considered
ways of coping with societal, relational, and personal matters in peopleÕs lives (Mintz
& Wright, 1993; cf. Brumberg, 2000). In this view, alcohol and food are both thought
to serve similar functions for an individual (Brisman & Seigel, 1984; cf. Snyder &
Cantor, 1998). The psychological structure of motivations in both domains might
have common features, even though the behaviors themselves are quite distinct
(cf. Cooper, Agocha, & Sheldon, 2000). Whereas the literature about psychological
motivations to eat is quite small (Arnow et al., 1995; Stunkard & Messick, 1985; van
Strien et al., 1986), there is considerable research on psychological motivations for
alcohol use.
Cox and Klinger (1988, 1990) proposed a framework for understanding psycho-
logical motivations for alcohol use, and this framework was applied by Cooper and
colleagues (Cooper, 1994; Cooper, Frone, Russell, & Mudar, 1995) to develop a
four-category model of motivation for alcohol use. As stated by Cooper (1994), in
this model, motivations for drinking alcohol
can be meaningfully characterized along two underlying dimensions reflecting the valence
(positive or negative) and source (internal or external) of the outcomes an individual hopes
to achieve by drinking. Thus, individuals may drink to obtain a positive outcome...or to
avoid a negative one... Moreover, drinking may be responsive to internal rewards, such
as the manipulation or management of oneÕs own internal emotional state, or to external
rewards, such as social acceptance or approval (p. 118).
Further, by crossing these two dimensions, four categories of motivations for al-
cohol consumption emerge. Specifically, we see a negatively valenced, internally elic-
ited motivation (to cope with negative affect); a positively valenced, externally
elicited motivation (to be social); a negatively valenced, externally elicited motivation
(to comply with social expectations); and a positively valenced, internally elicited
motivation (to enhance pleasure or positive affect).
The four-category model of drinking motivations has clear parallels to the types
of eating motivations discussed in the literature to date. For example, research has
shown that people regulate their eating behaviors as a way to cope with emotional
distress (Gangley, 1988; Heatherton & Baumeister, 1991; Stice & Agras, 1999; Strau-
man, Vookles, Berenstein, Chaiken, & Higgins, 1991). Whether emotional distress
leads to food restriction or intake depends on an interaction of numerous complex
factors: psychological, historical, cultural, and sociological (Bordo, 1993; Brumberg,
2000; Wolf, 1991). Both eating for social reasons and compliance reasons are influ-
enced by externally derived norms and expectations, though the motivation to eat to
be social is activated only during social occasions. Research suggests a norm of min-
imal eating in othersÕcompany is stronger than the norm of matching the food in-
take levels of oneÕs peers (Roth et al., 2001). Group norms, however, will shape the
baseline that constitutes minimal eating during social gatherings (e.g., in public set-
tings norms for young college women might lead to lesser degrees of food intake
than for their male counterparts; cf. Pliner & Chaiken, 1990; Wolf, 1991). The mo-
tivation to eat to comply—as activated by a set of internalized expectations which
may be chronically primed—can lead to either chronic food intake or restriction
(Fredrickson, Roberts, Noll, Quinn, & Twenge, 1998; Heatherton & Baumeister,
B. Jackson et al. / Journal of Research in Personality 37 (2003) 297–318 299
1991; Mintz & Betz, 1998; Noll & Fredrickson, 1998; Strauman et al., 1991; Wolf,
1991). Finally, eating for the sensory pleasure of food has been associated with
binge-eating. Indeed, sensory cues (e.g., delicious aroma or attractive appearance
of foods) have been shown to activate binge-eating episodes (cf. Waters, Hill, &
Waller, 2001), suggesting that pleasure motivations for eating might be the more
proximal cause of over-eating, mediating cue exposure and bingeing. Hence, al-
though the psychological motivations to eat thus far identified have not been inte-
grated into an overarching theoretical framework, they appear to fit well within
this framework (also see Cooper & Shapiro, 1997, and Cooper, Shapiro, & Powers,
1998, for an elaboration of this framework and its application to other health-
related behaviors).
In the current set of studies, we therefore develop and validate an inventory to
assess psychological motivations to eat based on this four-category model (i.e.,
coping, social, compliance, and pleasure motivations). In Study 1, possible scale
items were generated through an open-ended elicitation of 40 respondents, with ad-
ditional face valid items generated by the researchers. In Study 2, the factor struc-
ture of the inventory was determined, and item content was refined, based on data
collected from a separate sample of 812 college students. Given that gender funda-
mentally shapes eating practices and food consumption (Bordo, 1993; Wolf, 1991),
the factor structure was tested for invariance across females and males. As well,
measures to determine several types of validity—convergent, discriminant, and
concurrent—were included.
Based on theoretical and clinical knowledge of food consumption and of the eti-
ology of eating disorders, we formulated general hypotheses about how each of the
motivations would predict three types of eating behaviors: restrictive eating, binge-
ing, and purging. We hypothesized that (a) coping motivations to eat would posi-
tively predict all three types of eating behaviors; (b) social motivations to eat
would negatively predict all three types of eating behaviors; (c) compliance motiva-
tions to eat would positively predict all three types of eating behaviors; and (d) plea-
sure motivations to eat would positively predict binge-eating, negatively predict
restrictive eating, but show no relation to purging.
2. Study 1: Development of the item pool
In Study 1, we used open-ended elicitation procedures for three purposes: (1) to
determine whether the theoretically derived framework fits with peopleÕs own ideas
about reasons, beyond physiological needs, for eating; (2) to assess whether we might
be overlooking an important domain of these motivations; and (3) to obtain preli-
minary data on gender differences and similarities in these motivations. Eating dis-
orders are far more common among women than men, but are experienced by
both genders (Andersen, 1999; Wolf, 1991; cf. Bordo, 1993). By including both males
and females in the development of this instrument, we can explore whether the four
categories cover the major motivations to eat for both females and males, or if un-
ique domains exist for either gender.
300 B. Jackson et al. / Journal of Research in Personality 37 (2003) 297–318
2.1. Method
2.1.1. Sample and procedure
A convenience sample of 28 female and 12 male university and community volun-
teers responded in written form to open-ended questions about their motivations to
eat, beyond physical hunger. The mean age was approximately 25 years old. The fol-
lowing question was posed: ‘‘What kinds of feelings, thoughts, or circumstances typ-
ically prompt you (or others) to eat or want to eat (besides simply feeling hungry/
being hungry)?’’
2.2. Results and discussion
Participants generated 203 discrete responses to this question. Narrative responses
were independently coded by two raters into one of the four theoretically expected
motivations to eat: coping, social, compliance, and pleasure. Written descriptions
of each motivation domain were provided in advance to coders to ensure there
was a common understanding of each motivation. Responses judged as not fitting
these dimensions were assigned to an ‘‘other’’ category. Discrepancies between cod-
ers were resolved by discussion until consensus was reached. All psychologically
meaningful reasons for eating responses fit in one of the four categories. Of the re-
sponses generated, 48% were categorized as Coping motivations (e.g., ‘‘stress/de-
pressed’’), 13% were categorized as Social motivations (e.g., ‘‘as a social event’’),
4% were categorized as Compliance motivations (e.g., ‘‘someone saying Ôyou have
to try thisÕ’’), and 16% were categorized as Pleasure motivations (e.g., ‘‘love the taste
of food’’). The 19% of responses that were coded as other motivations were not cen-
trally psychological in nature (e.g., ‘‘hypoglycemic,’’ ‘‘keeping energy level up’’). In
sum, the four-category model seemed to fit well with participantsÕself-generated un-
derstandings about motivations to eat.
Both females and males generated responses consistent with the four-category
model. Of all the reasons for eating that females generated, 51% of the responses were
Coping motivations; 12% were Social motivations; 5% were Compliance motivations;
16% were Pleasure motivations; and 15% were Other motivations. Of all the reasons
for eating that males generated, 48% of the responses were Coping motivations; 15%
were Social motivations; 2% were Compliance motivations; 17% were Pleasure moti-
vations; and 29% were Other motivations. The differences by gender for percentage
generated in each category were not statistically different, v2ð4Þ¼6:00, p>:05.
3. Study 2: Development, refinement, and validation of the motivation to eat subscales
3.1. Method
3.1.1. Sample and procedure
An initial pool of items assessing each of the four theorized domains was gener-
ated from two sources. Thirty-three non-redundant responses were taken from Study
B. Jackson et al. / Journal of Research in Personality 37 (2003) 297–318 301
1 and edited as needed for clarity. Additional items were drawn from the literature to
supplement the Compliance, Social, and Pleasure categories. The resulting pool in-
cluded 39 items: 17 on the Coping subscale, 7 on the Social subscale, 7 on the Com-
pliance subscale, and 8 on the Pleasure subscale.
Participants for the present study were 812 undergraduates who received credit as
part of their introductory psychology course. The mean age of the sample was 18.8
years old, 64% of the sample was female, and 87% was European American.
3.1.2. Measures
Items were administered in random order, as part of a larger questionnaire
packet. Respondents rated the relative frequency of eating for each of the specified
reasons on a scale from 1 (almost never/never) to 5 (almost always/always).
To establish convergent and discriminant validity of the Motivations to Eat sub-
scales, the Emotional Eating Scale (EES; Arnow et al., 1995) and the Dutch Eating
Behavior Questionnaire (DEBQ; van Strien et al., 1986) were included in the ques-
tionnaire packet. The EES is a 25-item scale that assesses eating in response to neg-
ative emotions, and comprises three subscales assessing eating in response to anger/
frustration, anxiety, and depression, with 11, 9, and 5 items, respectively. Items begin
with the stem ‘‘How strong is your desire to eat when you feel...’’. Responses range
from 1 (no desire to eat)to5(an overwhelming urge to eat). In this study, the reliabil-
ity of each subscale was: Anger/Frustration, a¼:89; Depression, a¼:83; Anxiety,
a¼:83. These three subscales were expected to converge with the Coping subscale
in the Motivations to Eat measure. The DEBQ is a 33-item inventory that has three
subscales: Emotional Eating (eating in response to negative emotions), External Eat-
ing (eating in response to sensory cues), and Restrained Eating (how much one re-
strains or limits eating), with 13, 10, and 10 items, respectively. Responses range
from 1 (never)to5(very often). In this study, the reliability of each of the subscales
was: Emotional Eating, a¼:94; External Eating,
1
a¼:71; Restrained Eating,
a¼:91. Emotional Eating was expected to converge with the Coping subscale in
the Motivations to Eat measure. Though eating can be activated by sensory cues,
the pleasure-seeking motivation may be the more proximal cause of eating. As such,
External Eating was expected to demonstrate convergent validity with the Pleasure
subscale in the Motivations to Eat measure. Restrained Eating was expected to show
discriminant validity with each of the four dimensions of the Motivations to Eat
measure.
To assess the concurrent validity of the Motivations to Eat subscales, criterion
measures of three types of eating behaviors were also included. Respondents an-
swered questions about the presence and the lifetime frequency of eating restraint
(i.e., fasting, appetite control pill use, and strict dieting), bingeing, and purging
(i.e., vomiting, laxative use, and diuretic use). Presence was measured by a yes/no re-
sponse indicating whether the respondent had ever engaged in a given behavior. For
1
Due to a clerical error, seven items on the external eating subscale were missing from the
questionnaire administration.
302 B. Jackson et al. / Journal of Research in Personality 37 (2003) 297–318
purging and restraint, which were each assessed by multiple behaviors, a count of yes
responses was made across the relevant subset of behaviors. Thus scores could range
from 0 (no behaviors endorsed) to 3 (three behaviors endorsed) for both restrained
eating and purging. Respondents who had ever engaged in a given behavior were
then asked to indicate on a 1 (1–2 times in your life)to6(more than 20 times) scale
how often they had ever engaged in this behavior. For restrictive eating and purging,
lifetime frequency was determined by taking the average across the specific acts that
constituted the behavior.
Finally, to rule out the effects of socially desirable responding, we included the 20-
item version of the Marlowe–Crowne Social Desirability Scale (Strahan & Gerbasi,
1972). This scale comprises items that are answered true or false; half of the items are
reverse–scored. Sample items include ‘‘IÕm always willing to admit it when I made a
mistake’’ and ‘‘I like to gossip at times’’ (reverse–scored). In coding this scale, each
item was assigned the value of 0 or 1, and responses were then summed; 1 denotes
that the participant marked the socially desirable response. In the 1972 article, based
on a sample of 500 university students, the mean and standard deviation for this
scale are 14.5 and 5.4, respectively; no range was reported. In the current sample,
M¼12:21, SD ¼2:14, with answers ranging from 5 to 19.
3.2. Results and discussion
3.2.1. Overview of analyses
The analyses were conducted in five phases. First, the sample was randomly di-
vided into two subsamples (Group 1, n¼404; Group 2, n¼408).
2
The measurement
model was refined in Group 1, and the refined model was cross-validated in Group 2.
Means, standard deviations, and as of the resulting subscales were calculated. Sec-
ond, the final model was tested for invariance in the factor structure across gender,
using the pooled sample. Third, analyses were conducted to test for mean gender dif-
ferences in each of the four Motivations to Eat subscales. Fourth, convergent and
discriminant validity with selected established measures were examined. Fifth, multi-
ple regression was used to test the concurrent validity of the Motivations to Eat sub-
scales vis-
aa-vis three distinctive types of eating behavior: restrictive eating, bingeing,
and purging.
3.2.2. Measurement model refinement and cross-validation
Exploratory factor analyses. An exploratory factor analysis was conducted to de-
termine whether the four theorized factors would be extracted. Using data from
Group 1, we extracted factors through principal-axis factoring followed by oblimin
rotation, and examined the pattern matrix. Principal-axis factoring was used because
it has been shown to reproduce the initial correlation matrix more reliably than other
extraction techniques (Snook & Gorsuch, 1989), whereas oblimin rotation (which
2
Group 1 and Group 2 did not differ in gender composition or on mean levels of age, weight, socially
desirable responding, or any of the individual motivation to eat items.
B. Jackson et al. / Journal of Research in Personality 37 (2003) 297–318 303
allows correlated factors) was used because earlier research suggests that eating
motivations are correlated (e.g., van Strien et al., 1986). The resulting solution
was evaluated to determine whether items loaded on their intended factors at or
above .40, and demonstrated purity (Floyd & Widaman, 1995). In this case, items
whose loading on a secondary factor were less than half the value of the loading
on its primary or intended factor were considered pure (i.e., non-complex). For ex-
ample, if an item loaded. 60 on its primary factor and .30 or above on another factor,
it would be considered factorially complex.
Exploratory factor analysis of the 39 items yielded six factors with eigenvalues
greater than one. However, the sixth factor was barely over 1 (1.06), and the scree plot
suggested either 4 or 5 factors. Thus, the analysis was re-run constraining the solution
to five and four factors, respectively. In the five-factor solution, 18 eat to cope with
negative emotion items loaded on the first factor. All items loaded at or above .40,
and only two of the items failed to meet criteria for non-complexity. Six of the seven
social items loaded on factor two with loadings P:40. However, one of these six so-
cial items failed to meet our criterion for non-complexity, thus yielding five social
items that met both criteria. The third factor contained the compliance items, all
six of which met our dual criteria (i.e., loaded above .40 and showed purity). The
fourth factor was defined by four of the eight pleasure motivation items, with loadings
P:40. Three pleasure motivation items loaded on factor five, with loadings P:44.
One of the three pleasure items loading on factor five cross-loaded on factor four. Fi-
nally, one pleasure item loaded below .40 on both the fourth and fifth factors. In gen-
eral, the items loading on factors four and five were distinguished by the degree to
which the wording emphasized use of food as a reward or treat vs. eating in order
to experience pleasurable sensations. Constraining the solution to four factors yielded
a structure that was consistent with this conceptualization. Specifically, the first three
factors in the constrained four-factor solution were identical to those obtained in the
five-factor solution (reflecting Coping, Social, and Compliance factors, respectively),
whereas the seven items that had loaded on factors 4 and 5 combined to yield one
Pleasure motivation factor. Examination of individual item loadings on the four fac-
tors showed that all items loaded P:52 on their respective factors, and that all but
one item met our criteria for purity.
Based on the results of these analyses, a subset of five coping items that met our
dual criteria (i.e., loaded P:40 on its intended factor and had a 2:1 loading ratio on
the primary vs. secondary factors) and reflected a range of emotional states were se-
lected for inclusion in the final item pool. The five social items that met our dual cri-
teria were also retained, as were five of the six compliance items that met these
criteria. Finally, six of the seven pleasure items that met our dual criteria in the con-
strained four-factor solution were retained: three items reflecting motivations to eat
related to the pleasurable aspects of food, and three items reflecting the use of food
to reward or treat oneself. This ensured the broadest possible coverage of the plea-
sure motivations construct. Thus, a total of 21 items were retained on the four the-
oretically predicted factors.
Confirmatory factor analyses in Group 1. A correlated four-factor model was spec-
ified and tested in Group 1 using confirmatory factor analyses (CFA). The fit of the
304 B. Jackson et al. / Journal of Research in Personality 37 (2003) 297–318
model to the data was compared to three alternative models (see Table 1, models
1–4). A one-factor model tested the notion that a single underlying motivation is ad-
equate to account for the pattern of covariation among all of the items. Two alter-
native, correlated two-factor models were also tested: (a) we specified internal vs.
external factors in which Coping and Pleasure items were constrained to load on
the internal factor and Compliance and Social items were constrained to load on
the external factor; (b) we specified positive vs. negative reinforcement factors in
which Social and Pleasure items were constrained to load on the positive reinforce-
ment factor and Compliance and Coping items were constrained to load on the neg-
ative reinforcement factor.
All CFAs were performed using the EQS structural equation modeling program
(Bentler, 1995). The variance–covariance matrix served as input. Scaling metrics for
the latent variables were fixed by setting factor variances to 1.0. Following the rec-
ommendation of several authors (Bollen & Long, 1993; Loehlin, 1998), multiple fit
indices are reported for each of the analyses, including the v2/df ratio (Bollen,
1989), the normed fit index (NFI; Bentler & Bonnett, 1980), the comparative fix in-
dex (CFI; Bentler, 1995), and the standardized root-mean-square residual (RMR).
Given that each fit index has different limitations, consistency across indices is gen-
erally regarded as the most reliable indicator of goodness of fit (Mulaik et al., 1989).
Although no consensus exists on the exact value of the v2/df ratio needed to indicate
good fit, all recommendations fall in the range of 2:1 to no more than 5:1 (Bollen,
1989; Marsh & Hocevar, 1985). The values for both the NFI and the CFI range from
Table 1
Development and validation of factor structure across randomly split halves
Model v2df v2/df NFI CFI RMR
1. Original four-factor model: Group 1 749.51183 4.10 .88 .85 .05
2. Final four-factor model: Group 1 512.74164 3.13 .89 .92 .05
3. One-factor model: Group 1 1958.34170 11.52 .56 .59 .09
4. Two-factor modela
(internal vs. external): Group 1
1547.47169 9.16 .66 .68 .10
5. Two-factor modelb
(positive vs. negative): Group 1
1380.96169 8.17 .69 .72 .09
6. Final four-factor model: Group 2 354.77164 2.16 .92 .95 .04
7. Factor pattern invariant model:
Across groups
1195.56292 4.09 .91 .93 .05c
.04d
8. Factor-loading invariant model:
Across groups
898.35344 2.61 .90 .94 .05c
.04d
Note. NFI, Normed Fit Index; CFI, Comparative Fit Index; RMR, standardized root-mean-square
residual. In Group 1, n¼404, and in Group 2, n¼408 (total N¼812).
*
p<:001.
a
Compliance and Social items were constrained to load on an external factor; Coping and Pleasure
items were constrained to load on an internal factor.
b
Social and Pleasure items were constrained to load on a positive factor; Compliance and Pleasure
items were constrained to load on a negative factor.
c
Fit of the model to the data in Group 1.
d
Fit of the model to the data in Group 2.
B. Jackson et al. / Journal of Research in Personality 37 (2003) 297–318 305
0 to 1, with values from .90 to .95 as acceptable, and higher values indicating good fit
(Bentler & Bonnett, 1980; Bollen & Long, 1993; Hu & Bentler, 1999). The RMR is
the standardized average absolute difference between the original and the reproduced
matrices (Marsh & Hocevar, 1985). Small values (e.g., .05) indicate minimal discrep-
ancy between the original and the reproduced matrices. RMR values reported here
exclude diagonal elements and as such are more conservative.
Goodness-of-fit information for the four alternative models (tested in Group 1) is
summarized in Table 1, models 1–4. The original 21-item four-factor model did not
fit the data well, as demonstrated by NFI and CFI values of .88 and .85, respectively.
Examination of the modification indices (using the LaGrange Multiplier Test; Ben-
tler, 1995) revealed factorial complexity associated with the Pleasure item ‘‘because
eating is pleasurable and enjoyable.’’ For example, it appeared that there would be
substantial improvement of fit by allowing the error variance of this item to correlate
with other error variances. To construct a cleaner scale, and because the Pleasure
scale already had six items and the other scales had five, we re-specified the four-fac-
tor model by dropping this one item and proceeding with the five remaining items on
the Pleasure scale. This modified four-factor model yielded a more acceptable fit: the
v2/df ratio was 3.13, values for the NFI and CFI were .89 and .92, respectively, and
the RMR was .05.
As shown by the v2-difference test, the final correlated four-factor model fit the
data significantly better than did the one-factor model (Model 3–Model 2,
Dv2½6¼1445:60, p<:001), or either of the two-factor models (Model 4–Model 2,
Dv2½5¼1034:73, p<:001 and Model 5–Model 2, Dv2½5¼868:22, p<:001).
3.2.3. Validation in Group 2
To determine whether this modified factor structure was reliable, we specified and
tested the invariance of the correlated four-factor structure across the two random
samples. As shown in Table 1, line 6, the correlated four-factor model provided an
acceptably good fit to the data in the second random half of the sample. The v2/df ra-
tio was 2.16, values for the NFI and CFI were .92 and .95, respectively, and the RMR
was .04. To determine whether the specified model fit equally well across the two
groups, two simultaneous, between-group models were specified. First, a model
was tested in which a common factor-pattern was specified across Groups 1 and 2,
but the magnitude of the factor-loadings was allowed to vary. As shown on line 7
of Table 1, the v2/df ratio of 4.09, values for the NFI and CFI of .91 and .93, respec-
tively, and acceptable RMR values indicate that the specified four-factor model pro-
vided an equally adequate fit to the data across Groups 1 and 2, thus indicating
configural invariance across subsamples (Widaman & Reise, 1997).
Next, we tested a factor-loading equivalent model in which the factor-pattern as
well as the factor-loadings were constrained to equivalence across Groups 1 and 2
(line 8, Table 1). Examination of the fit indices for this model indicate that the fac-
tor-loading equivalent model also provided an acceptably good fit to the data across
subsamples. The v2/df ratio was 2.61, values for the NFI and CFI were .90 and .94,
respectively, with acceptable RMR values. However, the v2-difference test between
the factor-pattern only and the factor-pattern with factor-loading equivalent models
306 B. Jackson et al. / Journal of Research in Personality 37 (2003) 297–318
was significant, Dv2ð52Þ¼297:21, p<:01, indicating that the assumption of com-
plete factor-loading invariance was not supported.
When modification indices (using the LaGrange Multiplier Test) for individual
items were examined, results revealed factor-loading invariance for all individual
items across Groups 1 and 2 except for two items (one Coping item: ‘‘because you
feel worthless or inadequate’’ and one Compliance item: ‘‘because you donÕt want
to stand out or be different from others who are eating’’). The unstandardized load-
ing for the Coping item in Group 1 was :52 ð:041Þ,p<:001; in Group 2, it was
:64 ð:041Þ,p<:001. The unstandardized loading for the Compliance item in
Group 1 was 1:00 ð:067Þ,p<:001; in Group 2, it was :72 ð:059Þ,p<:001. De-
spite the lack of complete factor-loading invariance, all items loaded significantly
on the Coping motivations factor in both groups (tvalues ¼12.61–17.64 and
15.35–20.37 for Group 1 and Group 2, respectively, pvalues <:001), as did all Com-
pliance motivation items in both groups (tvalues ¼11.55–14.89 and 11.89–14.36 for
Group 1 and Group 2, respectively, pvalues <:001). In fact, all items loaded signif-
icantly in both groups for the other two scales as well (Social: tvalues ¼15.45–20.15
and 12.62–18.71 for Group 1 and Group 2, respectively, pvalues <:001; Pleasure: t
values ¼9.63–10.13 and 9.95–10.85 for Group 1 and Group 2, respectively, pvalues
<:001). Internal consistency was also comparable across groups for all four sub-
scales (Coping a¼:88 and .89; Compliance a¼:85 and .84; Social a¼:89 and
.88; Pleasure a¼:82 for both). Given strong evidence for configural invariance
and evidence of nearly complete factor-loading invariance across the subsamples
for the four-factor model, data were pooled for the remaining analyses.
Table 2 presents the standardized and unstandardized (with standard error) fac-
tor-loadings, for the correlated four-factor model estimated in the entire sample
(N¼812). Descriptive statistics and factor intercorrelations for the resulting scales
are presented in Table 3. As shown in Table 3, the four subscales possess adequate
reliability and show adequate independence among themselves. Interestingly, the
rank order of mean levels of endorsement for each motivation subscale is identical
to what has been observed in the alcohol motivations literature (Cooper, 1994). Spe-
cifically, Social (M¼2:72, SD ¼:93) and (Pleasure M¼2:31, SD ¼:81) motivations
are endorsed as the most common, followed by Coping (M¼1:65, SD ¼:78) and
Compliance (M¼1:58, SD ¼:70).
3.2.4. Invariance across gender
To determine whether the final correlated four-factor model was invariant across
females and males, two simultaneous, between-group models were specified. First, a
model was tested in which a common factor-pattern was specified across females and
males, but the magnitude of the factor-loadings was allowed to vary. As shown on
line 3 of Table 4, the v2/df ratio of 2.78, values for the NFI and CFI of .90 and
.93, respectively, and acceptable RMR values indicate that the specified four-factor
model provided an equally good fit to the data across females and males, thus indi-
cating configural invariance across subsamples.
Next, a factor-loading equivalent model was tested in which both the factor-pat-
tern and the factor-loadings were constrained to equivalence across females and
B. Jackson et al. / Journal of Research in Personality 37 (2003) 297–318 307
Table 2
Standardized and unstandardized factor loadings and standard errors for the hypothesized four-factor model using the pooled sample (N¼812)
Item Coping Social Compliance Pleasure
SFL USFL SE SFL USFL SE SFL USFL SE SFL USFL SE
9. Because youÕre depressed or sad .75 1.00 NAa
16. Because you feel worthless or inadequate .64 .58 .030
21. As a way to help you cope .88 .99 .037
22. As a way to comfort yourself .82 .94 .039
27. As a way to avoid thinking about something
unpleasant or to distract yourself
.74 .88 .041
7. Because itÕs a special or traditional part
of some social occasion or celebration
.70 .82 .041
29. As a way to enjoy a social gathering .82 .91 .033
31. As a way to celebrate a special occasion
with friends, family, or a loved one
.84 1.00 NAa
32. To be sociable .78 .82 .032
37. To join in a festive occasion .84 .95 .034
34. To keep people from asking questions
about why youÕre not eating
.74 1.00 NAa
35. Because someone pressures you to eat .75 .92 .046
36. Because you feel like you canÕt say ÔnoÕ.75 .95 .048
38. Because you dontwant to stand out or be
different from others who are eating
.79 .86 .045
39. To please your mother or someone else
who wants you to eat
.61 .91 .055
3. Because you want to treat yourself .76 1.00 NAa
5. As a reward for having done something
that youÕre proud of or feel good about
.72 .87 .058
6. Because you like to eat .53 .96 .045
10. Because you deserve it .79 .91 .045
19. Because you feel good or are in a good mood .69 .83 .044
Note. Items were preceded by the stem ‘‘How often do you eat.’’ Item numbers reflect the order in which items were presented to respondents. SFL,
standardized factor loading, USFL, unstandardized factor loading. All factor loadings are significant at p<:001.
a
NA, Not applicable. For identification purposes, the USFL was set to 1.
308 B. Jackson et al. / Journal of Research in Personality 37 (2003) 297–318
males. Examination of the fit indices for this model indicated that the factor-loading
equivalent model provided an acceptable fit to the data across females and males
(line 4, Table 4). Values for the NFI and CFI were .89 to .93, respectively, and
RMRs were acceptable (.05 for both groups). The v2-difference test was not signifi-
cant, Dv2ð16Þ¼1:77, p>:05, indicating that the assumption of complete factor-
loading invariance was supported. Nevertheless, examination of the modification
indices (using the LaGrange multiplier test) showed that relaxing the equality con-
straint on one Coping item (‘‘as a way to avoid thinking about something unpleasant
or to distract yourself’’) would significantly improve model fit. The unstandardized
loading for this item among females was :87 ð:052Þ,p<:001; among males, it was
1:00 ð:080Þ,p<:001. Not surprisingly, given earlier findings (see Table 2), all re-
maining items loaded significantly on their intended factors among both males
and females. As well, internal consistency was comparable across female and male
respondents on the Coping motivations subscale (a¼:88 and .87, respectively),
and on the Social (a¼:90 and .86, respectively), Compliance (a¼:85 and .83, re-
spectively), and Pleasure (a¼:82 and .83, respectively) motivations subscales.
These analyses indicate a largely, but not entirely, homogeneous solution. Social,
Compliance, and Pleasure motivations met the strict statistical assumptions of
Table 3
Scale statistics and factor correlations
Scale MSDaFactor correlations
1234
1. Coping 1.65 .78 .88 – .40 .48 .54
2. Social 2.72 .93 .88 .48 .65
3. Compliance 1.58 .70 .84 .41
4. Pleasure 2.31 .81 .82
Note. Data are from total sample (N¼812). Possible scale means range from 1 (almost never/never) to
5 (almost always/always), in response to item stems that began ‘‘How often do you eat....’’ Means,
standard deviations, and coefficient as are based on scale scores derived from manifest variables; inter-
correlations are among latent factors.
Table 4
Examination of factor invariance across gender
Model v2df v2/df NFI CFI RMR
1. Female (n¼520) 534.66164 3.26 .91 .93 .05
2. Male (n¼278) 360.04164 2.20 .88 .93 .05
3. Factor pattern invariant model 911.55328 2.78 .90 .93 .05a.05b
4. Factor-loading invariant model 913.27344 2.65 .89 .93 .05a.05b
5. v2difference 1.77 16 —
Note. NFI, Normed Fit Index; CFI, Comparative Fit Index; RMR, standardized root-mean- square
residual. Dashes indicate data not applicable.
*
p<:001.
a
Fit of the model to the data among Females.
b
Fit of the model to the data among Males.
B. Jackson et al. / Journal of Research in Personality 37 (2003) 297–318 309
complete factor-loading invariance for all items across gender groups, whereas all
but one Coping motivations item met the strict statistical assumptions of complete
factor-loading invariance. In summary, these data indicate factor-pattern invariance
for all scales across females and males and complete factor-loading invariance for all
but a single item. Given arguments that factor-pattern invariance is crucial, and that
more stringent forms of invariance cannot be reasonably expected between non-ran-
domly formed groups (Horn, McArdle, & Mason, 1983), our results suggest that
more than adequate similarities exist to permit meaningful cross-group comparisons.
3.2.5. Mean differences across gender in motivations to eat
To determine whether the Motivations to Eat subscales varied as a function of
gender, we conducted a one-way multivariate analysis of variance (MANOVA).
As shown in Table 5, results revealed that there were no gender differences in mean
levels of Social motivations (F¼:81, p>:10), Compliance motivations (F¼1:04,
p>:10), and Pleasure motivations (F¼1:55, p>:10). However, there were signif-
icant gender differences in Coping motivations (F¼24:60, p<:001). Female re-
spondents were significantly more likely than male respondents to endorse eating
to cope with emotional distress. Finally, supplementary analyses showed that this
difference was not due to age, weight, or socially desirable responding. Indeed, when
age, weight, and socially desirable responding were entered as covariates, the eta-
squared for Coping motivations increased, from about 5% of the variance to just
over 7%.
3.2.6. Convergent and discriminant validity
The data in Table 6 illustrate that the Motivations to Eat subscales correlated in
expected ways with the established measures included to assess convergent and dis-
criminant validity. The Coping motivations subscale showed the strongest significant
and positive correlations with each of the EES subscales (Anger/Frustration, Anxi-
ety, and Depression) and with the DEBQ Emotional Eating subscale, as expected.
Also as expected, Pleasure motivations were more strongly related to the DEBQ Ex-
ternal Eating Subscale than to any of the other scales. The modest correlations be-
tween Compliance motivations and the included criterion scales, as well as between
Social motivations and the criterion scales, indicate that neither Compliance nor So-
cial motivations are well-represented by existing measures. Finally, and as expected,
Table 5
Unweighted means for four motivations to eat subscales, by gender
Gender Coping Social Compliance Pleasure
M SD M SD SD SD M SD
Female (n¼520) 1.78 .82 2.74 .96 1.60 .73 2.33 .79
Male (n¼278) 1.40 .61 2.66 .89 1.54 .66 2.28 .84
g2.053.002 .001 .001
Note. WilksÕlambda ¼.94; multivariate F:ð4;807Þ¼13:81, p<:001.
*
p<:001.
310 B. Jackson et al. / Journal of Research in Personality 37 (2003) 297–318
the DEBQ Restrained Eating subscale correlated only modestly (rs<:30) with the
Motivations to Eat subscales, showing discriminant validity. Indeed, we would not
expect motivations to eat to be the same as motivations not to eat.
Finally, socially desirable responding did not correlate substantially with any of
the Motivations to Eat subscales, suggesting that responses to the subscales are more
than simply a function of what participants think they ought to say about their rea-
sons for eating. Overall, these data suggest that the eating motivations subscales
show both convergent and discriminant validity.
3.2.7. Motivations to eat as predictors of eating-related behavior
To establish concurrent validity of the Motivations to Eat subscales and the utility
of distinguishing among these motivations, we tested whether the subscales were dif-
ferentially related to three types of eating behaviors (restrictive eating, bingeing, and
purging). A series of six hierarchical multiple regression analyses was conducted. On
the first step, gender was entered and on the second step the set of four motivations to
eat was entered.
3
For each of the three types of eating behaviors the following two
indicators were examined as dependent variables: ever engaged in the behavior,
and lifetime frequency among those who ever engaged in the behavior. These indica-
tors were chosen so that we could examine the utility of the eating motivations sub-
scales in predicting both initial engagement and degree of involvement in the eating
behavior.
3
For dependent measures that were continuous, Ordinary Least Squares (OLS) estimation was used.
For the dichotomous dependent measure (ever binged), analyses were conducted with both logistic and
OLS regression procedures. Although the use of dichotomous dependent variables violates assumptions
underlying OLS estimation (see Neter, Kutner, Nachtsheim, & Wasserman, 1996), comparison of results
from the two procedures showed no substantive differences. Thus, to simplify reporting and maintain
comparability across dependent measures, results are tabled for the OLS regression procedures only.
Table 6
Correlations among motivations to eat subscales and convergent/divergent criterion scales
Coping Social Compliance Pleasure
Convergent/Divergent Criterion Scales
EES, Anger/Frustration Subscale 72 31 32 34
EES, Anxiety Subscale 67 29 34 34
EES, Depression Subscale 65 27 30 33
DEBQ, Emotional Eating Subscale 68 22 26 31
DEBQ, External Eating Subscale 34 33 17 47
DEBQ, Restrained Eating Subscale 30 08 21 03
M-C Social Desirability 08 02 05 12
Demographic variable
Gender (1 ¼male, 2 ¼female) 23 04 04 00
Note. N¼812, except n¼810 for Social Desirability. Except for Gender, all measures are scored so
that higher numbers indicate more of the measured construct. Decimals are omitted. All correlations
P:07 are significant at p<:05. EES, Emotional Eating Scale. DEBQ, Dutch Eating Behavior Ques-
tionnaire. MC, Marlowe–Crowne.
B. Jackson et al. / Journal of Research in Personality 37 (2003) 297–318 311
As shown in Table 7, the four Motivations to Eat subscales accounted for approx-
imately 3 to 13% of the variance across restrictive eating, bingeing, and purging. Ex-
amination of the individual beta weights indicated that each of the four motivations
was associated with a unique pattern of eating behavior.
Restrictive eating. As shown in Table 7, the four Motivations to Eat subscales ac-
counted for about five to six percent of the variance in restrictive eating, after ac-
counting for gender. In general, Coping and Compliance motivations each
significantly and positively predicted more restrictive eating for both indices. The ex-
ception was that Coping did not predict lifetime frequency of restrictive eating. Also
as expected, both Pleasure and Social motivations were significantly and negatively
related to more restrictive eating, though Pleasure did not predict lifetime frequency.
Binge eating. As shown in Table 7, the four Motivations to Eat subscales ac-
counted for about 13% of variance in the dichotomous binge eating measure, after
accounting for gender. Coping, Compliance, and Pleasure motivations were each sig-
nificantly and positively related to more binge eating. Social motivations were signif-
icantly and negatively related to more binge eating. Lifetime frequency of bingeing
was not related to any of the Motivations to Eat subscales. This suggests that the
motivations are primarily useful in distinguishing between those who ever vs. never
engaged in the behavior, and that they, at least in this sample, do not help us further
understand the degree of involvement in this behavior.
Purging. As shown in Table 7, the Motivations to Eat subscales accounted for
about three percent of the variance in the occurrence of purging behaviors, after ac-
Table 7
Multiple regression analyses predicting eating behaviors from four motivations to eat subscales
Variable DR2Coping Social Compliance Pleasure
Restrictive eatinga
Count ever (n¼811) .064 .15 ).09.24 ).05
Lifetime frequency
(n¼272)
.051 .16 ).25þ.42 .04
Bingeing
Ever binge (n¼810) .127 .14 ).05 .08 .08
Lifetime frequency
(n¼218)
.019 ).02 ).12 .19 .27
Purgingb
Count ever (n¼811) .031 .06).06.10 .01
Lifetime frequency
(n¼94)
.078 .01 .18 .40þ.13
Note. All analyses are after controlling for gender. Lifetime frequency variables include only those
respondents who have ever engaged in the behavior.
a
Each indicator of ‘‘Restrictive eating’’ is a composite measure of fasting, appetite control pill use, and
going on a strict diet.
b
Each indicator of ‘‘Purging’’ is a composite measure of vomiting, laxative use, and diuretic use.
+
p<:10.
*
p<:05.
**
p<:01.
***
p<:001.
312 B. Jackson et al. / Journal of Research in Personality 37 (2003) 297–318
counting for gender. Coping and Compliance motivations were significantly and pos-
itively related to purging, as expected. In contrast, but also as expected, Social moti-
vations were significantly and negatively related to purging, whereas Pleasure
motivations were unrelated. However, generally the motivations did not predict life-
time frequency of purging; Compliance did so only marginally so. As with bingeing,
this suggests that the motivations are primarily useful in distinguishing between those
who ever vs. never engaged in the behavior, and that they, at least in this sample, do
not really help us further understand the degree of involvement in this behavior.
3.2.8. Incremental validity
A series of regression equations was estimated to determine whether the four Mo-
tivations to Eat subscales would predict these criterion-related behaviors above and
beyond existing measures. Specifically, on the first step, the EES subscales, the
DEBQ subscales, and gender were entered, and on the second step, the set of four
Motivations to Eat subscales was entered. Because the Emotional Eating subscale
of the DEBQ and the EES subscales assess the same general content (the DEBQ
EE subscale correlates with EES anger/frustration, r¼:74; EES anxiety, r¼:67;
and EES depression, r¼:64, ps<:001), to reduce multicollinearity, only the EES
scales were included in the reported analyses.
4
As shown in Table 8, each motivation
to eat predicted the given eating behavior, in an overall pattern similar to in Table 7,
above and beyond the EES subscales and the DEBQ scale. There was one important
exception: including these measures on the first step eliminated or substantially re-
duced the predictive validity of Coping motivations. This was expected, though, gi-
ven the substantial overlap of the Coping motivations measure with the EES
subscales and DEBQ affect-related subscale noted in the convergent validity section.
Notably even with the DEBQ External Eating scale in the equation, Pleasure moti-
vations still predicted ever-bingeing. Thus, our measure appears to be an effective
shorter substitute for these longer measures.
In sum, each eating motivation independently is associated with a unique pattern
of eating behaviors, in expected ways, and predicts these behaviors above and be-
yond other established measures of eating motivations.
3.2.9. Effects of motivations to eat by gender
To determine if the effects of motivations to eat differ significantly between fe-
males and males, all six hierarchical multiple regression analyses from Table 7 were
estimated with the addition of Gender X Motivation interaction terms on the last
step. To reduce multicollinearity among the interaction terms and their constituent
variables, all variables were centered before computing and entering their interac-
tions (Aiken & West, 1991). To probe significant interactions, simple slopes were cal-
culated from the overall regression equation to illustrate the relationship between a
4
We also tested an alternative model in which the DEBQ Emotional Eating subscale was included
instead of the EES subscales. Results paralleled the findings reported in Table 8.
B. Jackson et al. / Journal of Research in Personality 37 (2003) 297–318 313
given eating motivation and criterion variable among females and males separately
(Aiken & West, 1991).
Six blocks of Gender X Motivation interaction terms (corresponding to each of
the six dependent measures in Table 7) were tested. Of these, the only significant in-
teraction was Coping differentially predicting lifetime frequency of binge eating by
gender. Specifically, for males but not for females, the stronger the Coping Motiva-
tion, the lower the lifetime frequency of binge eating in the full sample (for males,
b¼:29, p<:05; for females, b¼:11, p>:05). In sum, the analysis of interaction
effects suggests that eating motivation effects are largely invariant between females
and males, with a possible exception in predicting some aspects of binge eating.
4. General discussion
We developed and tested a four-category model of motivations to eat, extending
parallel research done in the domain of alcohol use motivations (Cooper, 1994). As
expected, analyses indicated the presence of four correlated factors that comprise
motivations to eat: Coping, Social, Compliance, and Pleasure. This model provided
an equally good fit for men and for women, as shown by tests of factorial invariance
and by demonstrating good reliability across gender groups. Analysis of mean differ-
ences across groups suggests that, while there are no gender differences in self-
Table 8
Multiple regression analyses predicting eating behaviors from four eating motivations, alternative analyses
Variable DR2Coping Social Compliance Pleasure
Restrictive eatinga
Count ever (n¼811) .037 .09 ).08.22 ).05
Lifetime frequency
(n¼272)
.034 .11 ).23þ.38 ).02
Bingeing
Ever binge (n¼810) .028 .07).05 .08 .05
Lifetime frequency
(n¼218)
.010 ).15 ).10 .20 .17
Purgingb
Count ever (n¼811) .016 .02 ).05.09 .01
Lifetime frequency
(n¼94)
.073 ).15 .27 .30 .21
Note. All analyses are after controlling for EES Anger/Frustration, EES Anxiety, EES Depression,
DEBQ External Eating, & Gender. Lifetime frequency variables include only those respondents who have
ever engaged in the behavior.
a
Each indicator of ‘‘Restrictive eating’’ is a composite measure of fasting, appetite control pill use, and
going on a strict diet.
b
Each indicator of ‘‘Purging’’ is a composite measure of vomiting, laxative use, and diuretic use.
+
p<:10.
*
p<:05.
**
p<:01.
***
p<:001.
314 B. Jackson et al. / Journal of Research in Personality 37 (2003) 297–318
reported mean levels of Social, Compliance, and Pleasure motivations to eat (even
when controlling for the potential confounds of age, weight, and socially desirable
responding), women endorse higher levels of Coping motivations than do men.
The validity of the correlated four-category model was further reinforced by the
fact that each eating motivation was associated with a unique pattern of eating be-
haviors, despite overlap among Motivations to Eat subscales. As hypothesized, Cop-
ing and Compliance motivations to eat both positively predicted restrictive eating,
bingeing, and purging. Given that these are both aversive motivations, we would ex-
pect some similarity between them. However, Coping and Compliance motivations
are differentiated from each other in that Coping motivations predicted bingeing
more strongly, and Compliance predicted restrictive eating and purging more
strongly (see Table 7). As expected, Pleasure motivations positively predicted binge
eating and negatively predicted restrictive eating, and showed no relation to purging.
Finally, Social motivations negatively predicted restrictive eating and purging, but
positively predicted bingeing.
It seems then that healthy eating is in part the result of meeting the needs of these
four psychological motivations in ways that are not dependent on food for their ful-
fillment, and that disturbed eating results from chronically acting on these motiva-
tions. For our present purposes, the absence of restrictive eating, bingeing, or
purging represents healthy eating. Thus, when a motivation negatively predicts a be-
havior, this points us toward an understanding of the psychological bases of healthy
eating. According to these data, social motivations negatively predict restrictive eat-
ing, bingeing, and purging. We speculate that healthy eating at least in part results
from thoughtful, creative, and flexible responses to social motivations, and disturbed
eating results from a unaware, chronic, and rigid responses to any of the psycholog-
ical motivations to eat (cf. Heatherton & Baumeister, 1991). Moreover, we suggest
that healthy eating requires that the needs corresponding to social motivations are
fulfilled in ways that are independent of food alone (Mintz & Wright, 1993; Roth,
1993). We acknowledge that we did not assess important aspects of healthy eating
behaviors, such as avoiding high fat or high sugar foods, or eating very healthy
foods, such as fruits and vegetables (Conner, Norman, & Bell, 2002). Assessing
how the four motivations predict these types of healthy eating behaviors could be
more fully explored in future research.
The Motivations to Eat measure is the first validated instrument that is both the-
oretically-based and that examines a range of psychological motivations pertaining
to eating. As expected, the Coping motivation subscale overlaps with other similar
established scales, like the DEBQ and EES subscales. Hence, if an instrument is
needed to distinguish among specific negative emotions, then researchers should con-
sider using measures such as the EES. If an instrument is needed that provides a
comprehensive measure of motivations that includes a more general Coping motiva-
tions subscale—akin to, but shorter than established scales like the DEBQ Emo-
tional Eating subscales—then our measure should be administered.
Such an instrument has both clinical and research utility. The Motivations to Eat
measure could be used in both individual treatment and prevention efforts of eating
disorders. The measure could be given to clients in treatment for eating disorders, as
B. Jackson et al. / Journal of Research in Personality 37 (2003) 297–318 315
well as those identified as at-risk for eating disorders, to assess the reasons underly-
ing their food intake. Counseling could then revolve around the identification of
other ways to meet these needs. The Motivations to Eat measure could also be used
in large-scale research aimed at identifying both risk and protective factors associ-
ated with the development of eating disorders. Studies could be conducted to exam-
ine the role that specific psychological motivations have in the etiology of eating
disorders, including bulimia, anorexia, and EDNOS (eating disorders not otherwise
specified). Likewise, it could be used in outcome research on the treatment of eating
disorders (i.e., do motivation scale scores change as a result of treatment?).
Several limitations of the present study should be acknowledged. First, our samples
were limited to college students. As well, while we speculate that our eating motivations
measure will have clinical relevance, we did not use indices of eating disorders per se,
nor did we examine the subscales in samples of individuals diagnosed with eating dis-
orders. Also, given the correlational nature of the data, we cannot make claims about
the temporal or causal relationships between the motivations and eating behaviors. Fi-
nally, we used only self-report data to measure the motivations and eating behaviors.
To begin to address these limitations, future research should collect psychometric
data on the entire 20-item measure using broader, more representative samples (e.g.,
across a different range of ages and cultural backgrounds). The subscales should also
be related to clinically-accepted indices of eating disorders, and administered to clin-
ical samples. Longitudinal designs examining the development of motivations to eat
and their causes and consequences would shed light on the temporal aspects we
could not address with the present data. Experimental designs would help us under-
stand the causal mechanisms involved with these motivations (e.g., can the priming
of individual motivations cause different types of eating behaviors?). Finally, as in
other research on motivations, non-self-report measures (e.g., projective measures,
behavioral observation) could be developed to assess motivations to eat.
In sum, this investigation extends knowledge about the psychological bases of eat-
ing behaviors, as prior research focused on only coping and compliance motivations
to eat. We do not intend this as necessarily a comprehensive model of all possible
reasons people eat, but of major psychological motivations for eating. As such, we
believe that in making parallels with motivational models in other health domains,
the Motivations to Eat measure is an important contribution.
Acknowledgments
This research was supported by Grant #AA08047 from the National Institute of
Alcohol Abuse and Alcoholism to M. Lynne Cooper. Thanks to Eileen Zurbriggen
for helpful comments regarding earlier drafts of this article.
References
Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions. Newbury
Park: Sage.
316 B. Jackson et al. / Journal of Research in Personality 37 (2003) 297–318
Allison, D. (1994). Handbook of assessment methods for eating behaviors and weight-related problems.
Thousand Oaks, CA: Sage.
Andersen, A. E. (1999). Gender-related aspects of eating disorders: A guide to practice. Journal of Gender-
Specific Medicine, 2, 47–54.
Arnow, B., Kenardy, J., & Agras, W. S. (1995). The Emotional Eating Scale: The development of a
measure to assess coping with negative affect by eating. International Journal of Eating Disorders, 18,
79–90.
Bentler, P. (1995). EQS Structural Equations Program manual. Encino, CA: Multivariate Software, Inc.
Bentler, P. M., & Bonnett, D. G. (1980). Significance tests and goodness-of-fit in the analysis of covariance
structures. Psychological Bulletin, 88, 588–600.
Bollen, K. A. (1989). Structural equations with latent variables. New York, NY: Wiley.
Bollen, K. A., & Long, J. S. (1993). Testing structural equation models. Newbury Park, CA: Sage.
Bordo, S. (1993). Unbearable weight: Feminism, Western culture, and the body. Berkeley: University of
California Press.
Brisman, J., & Seigel, M. C. (1984). Bulimia and alcoholism: Two sides of the same coin. Journal of
Substance Abuse Treatment, 1, 113–118.
Brumberg, J. J. (2000). Fasting girls: The history of anorexia nervosa. New York: Vintage Books.
Conner, M., Norman, P., & Bell, R. (2002). The theory of planned behavior and healthy eating. Health
Psychology, 21, 194–201.
Cooper, M. L., Agocha, V. B., & Sheldon, M. S. (2000). A motivational perspective on risky behaviors:
The role of personality and affect regulatory processes. Journal of Personality, 68, 1059–1088.
Cooper, M. L. (1994). Motivations for alcohol use among adolescents: Development and validation of a
four-factor model. Psychological Assessment, 6, 117–128.
Cooper, M. L., Frone, M. R., Russell, M., & Mudar, P. (1995). Drinking to regulate positive and negative
emotions: A motivational model of alcohol use. Journal of Personality and Social Psychology, 69, 990–
1005.
Cooper, M. L., & Shapiro, C. M. (1997). Motivations for health behaviors among adolescents. In J. A.
McNamara (Ed.), Creating the compliant patient (pp. 25–46). Ann Arbor, MI: Center for Human
Growth and Development, University of Michigan.
Cooper, M. L., Shapiro, C. M., & Powers, A. M. (1998). Motivations for sex and sexual behavior among
adolescents and young adults: A functional perspective. Journal of Personality and Social Psychology,
75, 1528–1558, Lead article.
Cox, M., & Klinger, E. (1988). A motivational model of alcohol use. Journal of Abnormal Psychology, 97,
168–180.
Cox, M., & Klinger, E. (1990). Incentive motivation, affective change, and alcohol use: A model. In M.
Cox (Ed.), Why people drink. New York: Gardner Press.
Floyd, F. J., & Widaman, K. F. (1995). Factor analysis in the development and refinement of clinical
assessment instruments. Psychological Assessment, 7, 286–299.
Fredrickson, B. L., Roberts, T.-A., Noll, S. M., Quinn, D. M., & Twenge, J. M. (1998). That swimsuit
becomes you: Sex differences in self-objectification, restrained eating, and math performance. Journal
of Personality and Social Psychology, 75, 269–284.
Gangley, R. M. (1988). Emotional eating and how it relates to dietary restraint, disinhibition, and
perceived hunger. International Journal of Eating Disorders, 7, 635–647.
Heatherton, T. F., & Baumeister, R. F. (1991). Binge eating as an escape from self-awareness.
Psychological Bulletin, 110, 86–108.
Horn, J. L., McArdle, J., & Mason, R. (1983). When is invariance not invariant: A practical scientistÕs
look at the ethereal concept of factor invariance. The Southern Psychologist, 1, 179–188.
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis:
Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1–55.
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.
Leham, A. K., & Rodin, J. (1989). Styles of self-nurturance and disordered eating. Journal of Consulting
and Clinical Psychology, 57, 117–122.
B. Jackson et al. / Journal of Research in Personality 37 (2003) 297–318 317
Loehlin, J. C. (1998). Latent variable models: An introduction to factor, path, and structural analysis (3rd
ed.). Mahwah, NJ: Erlbaum Associates.
Macht, M., & Simons, G. (2000). Emotions and eating in everyday life. Appetite, 35, 65–71.
Marsh, H. W., & Hocevar, D. (1985). Application of confirmatory factor analysis to the study of self-
concept: First- and higher-order factor models and their invariance across groups. Psychological
Bulletin, 97, 562–582.
Mintz, L. B., & Betz, N. E. (1998). Prevalence and correlates of eating disordered behaviors among
undergraduate women. Journal of Counseling Psychology, 35, 463–471.
Mintz, L. B., & Wright, D. (1993). Women and their bodies: Eating disorders and addictions. In E. Cook
(Ed.), Women, relationships, and power: Implications for counseling (pp. 211–246). Alexandria, VA:
American Counseling Association.
Mulaik, S. A., James, L. R., Van Alstine, J., Bennett, N., Lenni, S., & Stillwell, C. D. (1989). Evaluation of
goodness-of-fit indices for structural equation models. Psychological Bulletin, 105, 430–445.
Neter, J., Kutner, M. H., Nachtsheim, C. J., & Wasserman, W. (1996). Applied statistical models (4th ed.).
Chicago: Irwin.
Noll, S. M., & Fredrickson, B. L. (1998). A mediational model linking self-objectification, body shame,
and disordered eating. Psychology of Women Quarterly, 22, 623–636.
Pliner, P., & Chaiken, S. (1990). Eating, social motivations, and self-presentation in women and men.
Journal of Experimental Social Psychology, 26, 240–254.
Roth, D. A., Herman, C. P., Polivy, J., & Pliner, P. (2001). Self-presentational conflict in social eating
situations: A normative perspective. Appetite, 36, 165–171.
Roth, G. (1993). Breaking free from compulsive eating. New York: Plume Books.
Sherwood, N. E., Crowther, J. H., Wills, L., & Ben-Porath, Y. S. (2000). The perceived function of eating
for bulimic, subclinical bulimic, and non-eating disordered women. Behavior Therapy, 31, 777–793.
Snook, S. C., & Gorsuch, R. L. (1989). Component analysis versus common factor analysis: A Monte
Carlo Study. Psychological Bulletin, 106, 148–154.
Snyder, M., & Cantor, N. (1998). Understanding personality and social behavior: A functionalist strategy.
In D. T. Gilbert & S. T. Fiske (Eds.), The handbook of social psychology (Vol. 2, 4th ed., pp. 635–679).
New York: McGraw-Hill.
Steptoe, A., Pollard, T. M., & Wardle, J. (1995). Development of a measure of the motives underlying the
selection of food: The food choice questionnaire. Appetite, 25, 267–284.
Stice, E., & Agras, W. S. (1999). Subtyping bulimic women along dietary restraint and negative affective
dimensions. Journal of Consulting and Clinical Psychology, 67, 460–469.
Strahan, R., & Gerbasi, K. C. (1972). Short, homogenous versions of the Marlow-Crowne Social
Desirability Scale. Journal of Clinical Psychology, 28, 191–193.
Strauman, T. J., Vookles, J, Berenstein, V., Chaiken, S., & Higgins, E. T. (1991). Self-discrepancies and
vulnerability to body dissatisfaction and disordered eating. Journal of Personality and Social
Psychology, 61, 946–956.
Stunkard, A. J., & Messick, S. (1985). The three-factor eating questionnaire to measure dietary restraint,
disinhibition, and hunger. Journal of Psychosomatic Research, 29, 71–83.
Stunkard, A. J., & Messick, S. (1988). The Eating Inventory. San Antonio, TX: Psychological Corporation.
van Strien, T., Frijters, J. E. R., Bergers, G. P. A., & Defares, P. B. (1986). The Dutch Eating Behavior
Questionnaire (DEBQ) for Assessment of Restrained, Emotional, and External Eating. International
Journal of Eating Disorders, 5, 295–315.
Waters, A., Hill, A., & Waller, G. (2001). Internal and external antecedents of binge eating episodes in a
group of women with bulimia nervosa. International Journal of Eating Disorders, 29, 17–22.
Widaman, K. F., & Reise, S. P. (1997). Exploring the measurement in variance of psychological
instruments: applications in the substance abuse domain. In K. J. Bryant, M. T. Windle, & S. G. West
(Eds.), The science of prevention: methodological advances from alcohol and substance abuse research
(pp. 281–324). Washington, DC: American Psychological Association.
Wolf, N. (1991). The beauty myth: How images of beauty are used against women. New York: W. Morrow.
318 B. Jackson et al. / Journal of Research in Personality 37 (2003) 297–318
... However, they did not specifically examine eating in response or to cope with negative emotions. As eating in response or to cope with negative emotions has been shown to be especially prevalent in young adults (Barak et al., 2021) and associated with harmful outcomes, such as binging, purging, and restricting in undergraduate students (Jackson et al., 2003), this is an understudied but warranted construct to investigate in this population. ...
... The coping subscale of the Motivations to Eat (Jackson et al., 2003) measure was used to assess eating to cope. This measure is based on the same framework as the Drinking Motives Questionnaire-Revised (Cooper, 1994) and Marijuana Motives Measure (Simons et al., 1998), which has been used widely in the substance use literature, including the study by Wisener and Khoury (2021). 2 The coping subscale is comprised 1 Participants' height and weight were not collected as the relationships between dispositional mindfulness and self-compassion with eating pathology (including eating in response to negative emotions) have been shown to be independent of BMI (Gouveia et al., 2019;see Sala et al., 2020 for meta-analysis; see Turk & Waller, 2020 for meta-analysis). ...
... A sample item is, "As a way to comfort yourself." This measure has been validated in undergraduates (Jackson et al., 2003) and the coping subscale demonstrated excellent internal consistency (α = 0.902) in the present study. ...
Article
Eating in response or to cope with negative emotions has been shown to be problematic in undergraduate students. Dispositional mindfulness and self-compassion have been associated with less eating in response or to cope with negative emotions; however, specific mechanisms underlying these relationships have never been tested. The aim of the present study is to test whether lower levels of specific difficulties in emotion-regulation (i.e., non-acceptance of emotional responses, difficulties engaging in goal-directed behaviour, impulse control difficulties, lack of emotional awareness, limited access to emotion-regulation strategies, lack of emotional clarity) explain the negative relationship between dispositional mindfulness and self-compassion with eating to cope. Undergraduate students (N = 307) aged 18 to 24 (M = 20.28) completed online self-report measures. Dispositional mindfulness and self-compassion were associated with lower levels of non-acceptance of emotional responses, which in turn was associated with less eating to cope. Dispositional mindfulness and self-compassion were also associated with lower levels of difficulties engaging in goal-directed behaviour; however, difficulties engaging in goal-directed behaviour was unexpectedly associated with less eating to cope. To note, dispositional mindfulness was not associated with eating to cope when holding self-compassion constant. Results suggest that specific difficulties in emotion-regulation may explain the negative relationship between dispositional mindfulness and self-compassion with eating to cope. Findings contribute to theoretical models, and with replication can be used to inform the development of randomized-control trials examining the efficacy of mindfulness and self-compassion-based training for eating to cope in undergraduate students.
... Crossing these two dimensions creates four distinct motives for alcohol use: self-focused approach (to enhance positive affect), self-focused avoidance (to reduce negative affect), social-focused approach (to bond or connect with others), and social-focused avoidance (to conform with social expectations or pressure; Cooper, 1994). Jackson et al. (2003) adapted this model to describe four motives for both healthy and disordered eating: enhancement (to experience pleasure), coping (to reduce negative affect), social (to bond or connect with others), and compliance (to comply with social expectations or pressure). Similarly, the four-function model of NSSI posits two orthogonal dimensions (automatic versus social; positive reinforcement versus negative reinforcement) that when crossed, create four motives for NSSI: automatic positive reinforcement (to generate desirable internal states), automatic negative reinforcement (to reduce undesirable internal states), social positive reinforcement (to elicit desirable social states), and social negative reinforcement (to decrease undesirable social states; Nock & Prinstein, 2004). ...
... Although these motivational models appear strikingly similar, a closer examination of the items typically used to measure each dimension reveals two key differences between them. First, whereas the alcohol use and disordered eating models include affiliative motives related to both bonding (e.g., "To be sociable") and conforming (e.g., "To fit in with a group you like"; Cooper, 1994;Jackson et al., 2003), the affiliative motives within the four-function model of NSSI focus exclusively on conforming (e.g., "To feel more a part of a group"; Nock & Prinstein, 2004). Second, the four-function model of NSSI incorporates communicative motives (e.g., "To let others know how desperate you are"; Nock & Prinstein, 2004), whereas the alcohol use and disordered eating models do not (Cooper, 1994;Jackson et al., 2003). ...
... First, whereas the alcohol use and disordered eating models include affiliative motives related to both bonding (e.g., "To be sociable") and conforming (e.g., "To fit in with a group you like"; Cooper, 1994;Jackson et al., 2003), the affiliative motives within the four-function model of NSSI focus exclusively on conforming (e.g., "To feel more a part of a group"; Nock & Prinstein, 2004). Second, the four-function model of NSSI incorporates communicative motives (e.g., "To let others know how desperate you are"; Nock & Prinstein, 2004), whereas the alcohol use and disordered eating models do not (Cooper, 1994;Jackson et al., 2003). One explanation for these differences could be that bonding and communicative motives are not relevant to certain SDBs. ...
Article
Integrating across motivational models suggests that different self-damaging behaviors (SDBs) are enacted for similar reasons. However, it remains unclear whether some motives are more relevant to certain SDBs than others. To answer this question, the present study compared the salience of eight potentially shared motives across three exemplar SDBs, selected to represent different points along the internalizing and externalizing spectra: binge drinking, disordered eating (binge eating, purging, fasting), and nonsuicidal self-injury (NSSI). 704 first-year university students (73% female, Mage = 17.97) completed monthly surveys assessing their engagement in and motives for SDBs. Motives were conceptualized as either interpersonal (bonding with others, conforming with others, communicating strength, communicating distress, reducing demands) or intrapersonal (reducing negative emotions, enhancing positive emotions, punishing oneself). Multilevel models compared endorsement of each motive across SDBs. Results revealed that SDBs were motivated by similar goals, albeit to different degrees. Although some exceptions emerged, interpersonal motives were most salient to binge drinking, followed by disordered eating, and then NSSI. In contrast, intrapersonal motives were most salient to NSSI, followed by disordered eating, and then binge drinking. Motivational differences were also found within disordered eating. For example, punishing oneself was more relevant to purging and fasting than binge eating, whereas relieving negative emotions was more relevant to binge eating and purging than fasting. Similar to dimensional models that position SDBs on internalizing or externalizing spectra, the salience of motives for binge drinking and NSSI may fall on distinct spectra (i.e., interpersonal and intrapersonal, respectively), with motives for disordered eating exhibiting elements consistent with both spectra. This study supports a common motivational framework for investigating and potentially treating a variety of topographically distinct SDBs.
... Nutrition Reviews V R Vol. 00(0): [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17] in insulin level across the entire misalignment cycle, indicating a possible exaggerated postprandial glucose response. Moreover, 37% of study participants exposed to circadian misalignment had meal responses consistent with a prediabetic or diabetic state. ...
Article
Context Recent studies show that dietary habits and obesity seem to be influenced by chronotype, which reflects an individual’s preference for the timing of sleeping, eating, and activity in a 24-hour period. Objective This review aimed to analyze the association of chronotype with dietary habits, namely energy and macronutrient intakes, meal timing, and eating patterns, as well as with obesity. Data Sources PubMed/MEDLINE, LILACS, and Google Scholar databases were searched between 2004 and 2020. Study selection was performed by 2 authors independently; disagreements on eligibility of articles were resolved by a third author. After assessment of 12 060 abstracts, 43 studies (21 articles on obesity; 13 on food consumption, meal timing, and eating patterns; and 9 that addressed both obesity and dietary behavior) were included. Data Extraction A standard form was used to extract study design, country, number of participants, method of chronotype determination, and main findings. Data Analysis Approximately 95% of included studies showed an association between eveningness and at least 1 unhealthy eating habit. Morningness was associated with regular consumption of fresh and minimally processed foods. In addition, about 47% of studies showed a higher association between late types and obesity. Conclusion Late types are more likely to present unhealthy eating habits, such as eating late at night, skipping breakfast often, and eating processed/ultraprocessed foods, while early types are more likely to have healthy and protective habits, such as eating early and eating predominantly fresh/minimally processed foods. Intermediate types tend to have a pattern of health and eating more similar to early types than to late types. Late types are also more likely to present higher weight and body mass index than early or intermediate types. Systematic Review Registration PROSPERO registration no. CRD42021256078.
Article
Objectives: This study examines motivations for yoga and identifies unique motivational profiles among a sample of young adult yoga practitioners. This study further determines how young adult yoga practitioners' motivational profiles associate with physical health behaviors and psychological factors. Subjects/Setting: Survey data were drawn from the fourth wave of a large, population-based study (Project EAT-IV; Eating and Activity in Teens and Young Adults). Design: Latent class analysis (LCA) was used to identify motivational profiles among Project EAT-IV participants practicing yoga (n = 297; mean age: 30.8-1.7 years; 79.7 % female). Cross-sectional associations between latent motivational profiles, physical health behaviors, and psychological factors were determined with unadjusted and adjusted (gender, race/ethnicity, and body mass index) general linear models. Results: Across motivational profiles, most young adult yoga practitioners were motivated by enhanced fitness and stress reduction/relaxation. Additional motivations for yoga clustered by appearance (desire to change body appearance or weight) or mindfulness (desire to increase present moment awareness) underpinnings. The LCA characterized motivational profiles as "Low Appearance, Low Mindfulness" (Class 1; n = 77), "Low Appearance, High Mindfulness" (Class 2; n = 48), "High Appearance, Low Mindfulness" (Class 3; n = 79), and "High Appearance, High Mindfulness" (Class 4; n = 93). Having a profile with high mindfulness and low appearance motivations (Class 2) was associated with higher body satisfaction in comparison to the other classes (p < 0.001). Relative to Class 2, those with low mindfulness motivations (Class 1; Class 3) reported less total physical activity (p = 0.002) and those with high appearance motivations (Class 3; Class 4) reported higher compulsive exercise scores (p = 0.002). Conclusions: In this sample, high mindfulness and low appearance motivations for yoga appeared optimal for physical and psychological health. Cross-sectional findings suggest that young adult yoga practitioners' mind-body health may be supported by motivational underpinnings that emphasize yoga's internal (mindfulness) rather than external (appearance) benefits.
Article
Full-text available
Nutritional warnings (NWs) as a front-of-package label were implemented as a public policy aiding consumers with recognizing processed foods with high levels of critical nutrients (sodium, saturated fats, carbohydrates, and calories). However, in spite of this tool being well positioned in consumer decision making, there is little extant knowledge about the relationship between the message sent by NW, nutritional knowledge, consumer motivation, and the intention to avoid consuming processed foods. To understand these dimensions’ relations, a theoretical model was created and subsequently tested through structural equations. We applied a survey to 807 home food purchasing decision makers. The results show that the direct effect of NW messages raises the intention to avoid processed foods, while eating motivation is negative in its direct effect on the same avoidance intention. However, the message sent by NWs had a mediating effect between the intentions to avoid processed food and eating motivation but showed no such effect on nutritional knowledge. This suggests that the message sent by NWs was able to turn negative eating motivation into positive eating motivation to avoid processed foods. In conclusion, NWs help mitigate eating motivations, as well as boost the intention to avoid processed foods.
Article
Full-text available
Weight-based bullying is among the most prevalent forms of peer harassment and carries significant health consequences—particularly among adolescents who identify as a sexual and/or gender minority (SGM). We examined how anti-bullying legislation that includes weight as a protected class (enumeration) contributes to the prevalence of weight-based bullying and its adverse health sequelae among SGM adolescents. We collected data on weight-based bullying and health risk (stress, unhealthy eating behaviors, self-rated health) from the LGBTQ National Teen Survey and linked these to state anti-bullying legislation obtained from the United States (US) Department of Health and Human Services. Weight-based bullying was less frequent, but associated with greater health risk for SGM adolescents in states with, versus without, weight-enumerated anti-bullying laws. Adding weight as a protected class in anti-bullying legislation may be an effective strategy for reducing weight-based bullying, but additional supports are needed to support adolescents who continue to experience weight-based bullying.
Article
Mindful eating behavior and self-compassion have been individually shown to contribute to healthy eating. Evidence suggests that they may interact and that interaction may increase our understanding of eating behavior. The aim of the current study was to explore the relationship between mindful eating behavior, self-compassion, healthy/unhealthy weight, eating disorder and wellbeing in a sample of university students. An online survey using questionnaire data collection in a sample of 349 students (105 males and 244 females). Variables measured included body mass index (BMI), mindful eating behavior, eating attitudes, self-compassion and mental wellbeing. Regression and path analysis show mindful eating behavior and self-compassion to individually and interaction predict eating attitudes and wellbeing. BMI was shown to have a curvilinear relationship with mindful eating behavior, self-compassion and wellbeing with both the underweight and obese reporting lower self-compassion, less Mindful eating behavior, lower wellbeing and more likely to exhibit disordered eating. These results indicate that self-compassion and mindful eating behavior might be usefully targeted in interventions to prevent both obesity and eating disorder.
Article
The aim of this study is to identify the prevalence, predictors, and psychosocial well-being (depressive symptoms, stress) and weight-related behavior (eating behaviors, physical activity, and sedentary behavior) correlates of social distancing during COVID-19 among emerging adults. A rapid-response survey was sent to participants ( n=720; mean age=24.7 ± 2.0 years, 62% female) in a population-based cohort study in Minnesota during April–October 2020. Half of emerging adults reported fully social distancing. Emerging adults from White backgrounds were least likely to social distance while those from Asian backgrounds were most likely to social distance, in addition to those living with a parent. Females who partially/did not social distance reported less healthy eating behaviors, while males and “essential workers” reported higher levels of psychosocial distress. Public health messaging and practical supports for social distancing may need to be made more relevant to emerging adults during public health crises. Resources may need to differ depending on sex of emerging adult.
Article
There has been little recognition that people with type 2 diabetes are vulnerable to weight stigma and diabetes stigma and almost no research examining the implications of these forms of stigma for their health and well-being. This study examined health behavior correlates of weight stigma and diabetes stigma in 1,227 adults with type 2 diabetes. Results showed that experiencing weight stigma in health care, experiencing differential treatment from others because of their diabetes, and engaging in self-stigma for diabetes and body weight were each significantly associated with increased frequency of binge eating and eating as a coping strategy to deal with negative feelings. Internalizing weight stigma was also significantly associated with lower levels of physical activity and worse self-rated health. These findings suggest that initiatives to improve the health and well-being of people with type 2 diabetes must consider the potentially harmful roles of weight stigma and diabetes stigma.
Article
The development of the Emotional Eating Scale (EES) is described. The factor solution replicated the scale's construction, revealing Anger/Frustration, Anxiety, and Depression subscales. All three subscales correlated highly with measures of binge eating, providing evidence of construct validity. None of the EES subscales correlated significantly with general measures of psychopathology. With few exceptions, changes in EES subscales correlated with treatment‐related changes in binge eating. In support of the measure's discriminant efficiency, when compared with obese binge eaters, subscale scores of a sample of anxiety‐disordered patients were significantly lower. Lack of correlation between a measure of cognitive restraint and EES subscales suggests that emotional eating may precipitate binge episodes among the obese independent of the level of restraint. The 25‐item scale is presented in an Appendix (Arnow, B., Kenardy, J., & Agras, W.S.: International Journal of Eating Disorders, 17, 00‐00, 1995). © 1995 by John Wiley & Sons, Inc.
Article
The development of the Dutch Eating Behaviour Questionnaire (DEBQ) with scales for restrained, emotional, and external eating is described. Factor analyses have shown that all items on restrained and external eating each have high loadings on one factor, but items on emotional eating have two dimensions, one dealing with eating in response to diffuse emotions, and the other with eating in response to clearly labelled emotions. The pattern of corrected item-total correlation coefficients and of the factors was very similar for various subsamples, which indicates a high degree of stability of dimensions on the eating behavior scales. The norms and Cronbach's alpha coefficients of the scales and also the Pearson's correlation coefficients to assess interrelationships between scales indicate that the scales have a high internal consistency and factorial validity. However, their external validity has yet to be investigated.