Motivational "Spill-Over" During Weight Control: Increased Self-Determination and Exercise Intrinsic Motivation Predict Eating Self-Regulation

Abstract

Successful weight management relies on at least two health behaviors, eating and exercise. However, little is known about their interaction on a motivational and behavioral level. Based on the Hierarchical Model of Motivation the authors examined whether exercise-specific motivation can transfer to eating regulation during a lifestyle weight control program. The authors further investigated whether general, treatment-related, and exercise motivation underlie the relation between increased exercise and improved eating regulation.
Overweight/obese women participated in a 1-year randomized controlled trial (N = 239). The intervention focused on promoting physical activity and internal motivation for exercise and weight loss, following Self-Determination Theory. The control group received general health education.
General and exercise specific self-determination, eating self-regulation variables, and physical activity behavior.
General self-determination and more autonomous exercise motivation predicted eating self-regulation over 12 months. Additionally, general and exercise self-determination fully mediated the relation between physical activity and eating self-regulation.
Increased general self-determination and exercise motivation seem to facilitate improvements in eating self-regulation during weight control in women. These motivational mechanisms also underlie the relationship between improvements in exercise behavior and eating regulation.

Full text

APA PROOFS
Motivational “Spill-Over” During Weight Control:
Increased Self-Determination and Exercise Intrinsic
Motivation Predict Eating Self-Regulation
Jutta Mata, Marlene N. Silva, Paulo N. Vieira, Eliana V. Carrac¸a, Ana M. Andrade, Sı´lvia R. Coutinho,
Luis B. Sardinha, and Pedro J. Teixeira
Technical University of Lisbon
Objective: Successful weight management relies on at least two health behaviors, eating and exercise.
However, little is known about their interaction on a motivational and behavioral level. Based on the
Hierarchical Model of Motivation the authors examined whether exercise-specific motivation can
transfer to eating regulation during a lifestyle weight control program. The authors further investigated
whether general, treatment-related, and exercise motivation underlie the relation between increased
exercise and improved eating regulation. Design: Overweight/obese women participated in a 1-year
randomized controlled trial (N⫽239). The intervention focused on promoting physical activity and
internal motivation for exercise and weight loss, following Self-Determination Theory. The control group
received general health education. Main Outcome Measures: General and exercise specific self-
determination, eating self-regulation variables, and physical activity behavior. Results: General self-
determination and more autonomous exercise motivation predicted eating self-regulation over 12 months.
Additionally, general and exercise self-determination fully mediated the relation between physical
activity and eating self-regulation. Conclusion: Increased general self-determination and exercise mo-
tivation seem to facilitate improvements in eating self-regulation during weight control in women. These
motivational mechanisms also underlie the relationship between improvements in exercise behavior and
eating regulation.
Keywords: Multibehavior change, autonomy, obesity, randomized controlled trial, physical activity
In contrast to many other health-enhancing treatments, weight
management programs almost always target changes in two dif-
ferent behaviors: eating and physical activity. These two behaviors
tend to cluster in cross-sectional studies (Pronk et al., 2004) and
may also display interactive effects in intervention studies (C. L.
Dunn et al., 2006; Jakicic, Wing, & Winters-Hart, 2002). Baker
and Brownell (2000) suggested that exercise may play a key role
in long-term weight management by influencing both physiologi-
cal processes such as energy metabolism and appetite (see also
Martins, Morgan, & Truby, 2008 for a review), as well as psycho-
logical aspects like self-efficacy, body image, or mood. Baker and
Brownell argued it was important that the latter mechanisms might
also result in stronger motivation and confidence, which would in
turn improve eating self-regulation leading to better dietary com-
pliance (as well as long-term exercise adherence). This model has
since been partially tested and supported (Annesi & Unruh, 2008).
Yet another pathway by which exercise might positively affect
the regulation of eating behavior is through its influence on vari-
ables such as motivation, commitment, and feelings of efficacy
(Baker & Brownell, 2000). These effects could involve both quan-
titative and qualitative dimensions. On the one hand, success in
adopting an exercise plan could increase confidence (self-
efficacy), internal locus of control, and the overall motivational
drive toward other behaviors involved in weight management,
such as restricting energy-dense foods, self-monitoring, and adopt-
ing stress management practices. At the same time, it is possible
that this motivational “spill-over effect” could also depend upon
the quality of the motivation involved, specifically whether the
exercise motivation is characterized by an internal locus of cau-
sality, more intrinsic motives to be active, and fueled by feelings
of autonomy and self-determination (high sense of volition), as
opposed to motivation being externally driven, such as to please
others, and subject to strong controlling influences (lower au-
tonomy and volition).
More autonomous and intrinsic motivation have been shown to
be powerful predictors of successful self-regulation in the domains
of exercise (Fortier, Sweet, O’Sullivan, & Williams, 2007), eating
(Pelletier & Dion, 2007; Pelletier, Dion, Slovinec-D’Angelo, &
Reid, 2004), weight loss, and weight loss maintenance (Teixeira et
al., 2006; Williams, Grow, Freedman, Ryan, & Deci, 1996). How-
Jutta Mata, Marlene N. Silva, Paulo N. Vieira, Eliana V. Carrac¸a, Ana
M. Andrade, Sı´lvia R. Coutinho, Luis B. Sardinha, and Pedro J. Teixeira,
Department of Exercise and Health, Faculty of Human Kinetics, Technical
University of Lisbon, Cruz Quebrada, Portugal.
This study was partially funded by the Portuguese Science and Tech-
nology Foundation grants POCI/DES/57705/2004 and SFRH/BPD/35953/
2007 (to Jutta Mata), and by the Calouste Gulbenkian Foundation grant
65565/2004. The authors thank Teresa Santos and Mariana Pessoa for their
help with this study.
Correspondence concerning this article should be addressed to Dr. Jutta
Mata, Department of Exercise and Health, Faculty of Human Kinetics,
Technical University of Lisbon, Estrada da Costa, 1495-688 Cruz Que-
brada, Portugal. E-mail: jmata@fmh.utl.pt
Health Psychology © 2009 American Psychological Association
2009, Vol. ●●, No. ●, 000– 000 0278-6133/09/$12.00 DOI: 10.1037/a0016764
1
AQ: 1
tapraid5/zg1-hea/zg1-hea/zg100509/zg12352d09z
xppws S⫽1 7/8/09 22:47 Art: 2009-1642

APA PROOFS
ever, it is unclear if affecting self-determination in one health
domain has repercussions on the motivational regulation for other
health behaviors. The hierarchical model of motivation (Vallerand,
1997; Vallerand & Ratelle, 2002) predicts that such a motivational
transference is possible. It suggests that motivation operates at
three hierarchically ordered levels, the situational, contextual, and
global level. Situational motivation relates to a specific scenario,
for example a run on a Saturday morning. Contextual motivation
refers to a specific life context or domain, such as physical activity.
Global motivation is the most general construct, akin to a person-
ality construct, such as whether a person’s motivation is generally
more internally or more extrinsically oriented. The three levels
dynamically influence each other through both top-down and
bottom-up processes. Top-down processing refers to the impact of
motivation of a higher level on a lower level. For example, if a
person is generally self-determined toward physical activity, she
will likely feel self-determined while engaging in a specific exer-
cise activity that is relevant to her. Bottom-up processes occur
when experiences on a lower level affect motivation at a higher
level; for example, repeated experiences of autonomy and strong
volition in specific exercise situations might affect contextual
motivation toward physical activities in general, which would
eventually contribute to a more self-determined motivation style
(Vallerand, 1997).
A number of observational studies have shown how different
levels of motivation affect each other and also how they affect
behavior. Both bottom-up and top-down processes in motivation
were shown in the domains of exercise between all three levels of
motivation (Blanchard, Mask, Vallerand, de la Sablionnie`re, &
Provencher, 2007). Pelletier and colleagues (Pelletier et al., 2004)
reported an association between global self-determined orientation
at baseline and eating-specific self-determination at follow-up 13
weeks later. Also, a positive relationship between a general level
of self-determination and autonomous regulation of eating behav-
ior was shown (Pelletier & Dion, 2007). In how far global self-
determination directly affects behavior was studied in a quasi-
experiment by Williams and colleagues (Williams et al., 1996):
They found that general and treatment autonomy orientation pre-
dicted both attendance to a weight loss program and actual weight
loss in a 6-month weight loss intervention, and also weight loss
maintenance and exercise behavior at 2-year follow up.
Another important aspect of motivation transfer between two or
more behaviors is the different contexts in which they might occur.
For example, eating likely occurs in a family or work-related
setting, whereas physical activity within a weight loss program
could occur in exercise classes, group activities, or individual
leisure time. Hagger and colleagues (Hagger, Chatzisarantis, Cul-
verhouse, & Biddle, 2003) suggest that motivation underlying one
behavior can transfer from one context to the next (“trans-
contextual model”). Specifically, they showed that perceived au-
tonomy support and intrinsic motivation in the context of physical
education classes affects leisure time physical activity locus of
causality and identified regulation. These findings have been rep-
licated cross-culturally, showing that perceived autonomy support
and autonomous exercise motives in physical education class at
school transfer at least partially to exercise motivation in leisure
time activities (Hagger, Chatzisarantis, Barkoukis, Wang, & Bara-
nowski, 2005). To our knowledge, the dynamic interplay between
the different contexts and hierarchical levels of motivation across
different behaviors has not been tested in an experimental weight
control trial.
Goals
We sought to investigate how general, treatment, and exercise-
specific self-determined motivation relate to markers of eating
self-regulation in the context of a weight management program.
Specifically, we hypothesized that i) general, treatment, and
exercise-specific self-determination and motivation transfer to,
that is, are associated with important markers of eating self-
regulation, and that ii) self-reported physical activity is associated
with eating self-regulation through its effects on (i.e., mediated by)
general self-determination, treatment motivation, as well as
exercise-specific motivation.
Method
Design
The study was a randomized controlled trial in overweight and
moderately obese women, primarily focused on increasing exer-
cise self-motivation and exercise adherence, aiming at long-term
weight control. The intervention group participated in weekly or
biweekly sessions for approximately 1 year. Intervention targets
included increasing physical activity and energy expenditure,
adopting a diet consistent with a moderate energy deficit, and
ultimately establishing exercise and eating patterns that would
support weight maintenance. The program’s principles and style of
intervention were based on Self-Determination Theory (Deci &
Ryan, 1985; Ryan & Deci, 2000) and focusing on increasing
efficacy and self-determination toward exercise and weight con-
trol, while supporting participants’ autonomous decisions as to
which changes they wanted to implement and how. The control
group received a general health education program. The interven-
tion and its theoretical rationale have been described in detail
elsewhere (Silva et al., 2008). The Faculty of Human Kinetics
Ethics Committee reviewed and approved the study.
Participants
Participants were recruited from the community at large through
media advertisements. By design, only premenopausal women
(N⫽258) were accepted into the study. Of these, 19 women were
subsequently excluded from all analyses because they started
taking medication (e.g., antidepressants, anxiolytics, antiepilep-
tics) susceptible to affect weight (n⫽10), had a serious chronic
disease diagnosis or severe illness/injury (n⫽4), became pregnant
(n⫽2) or entered menopause (n⫽3). These 19 women were of
similar age ( p⫽.58) and Body Mass Index (BMI; p⫽.42) as the
239 participants considered as the valid initial sample.
Participants were between 23 and 50 years old (38 ⫾6.8 years)
and were overweight or mildly obese, with an initial BMI of
31.3 ⫾4.1 kg/m
2
. They were relatively well educated: 67% had at
least some college education, 23% had between 10 and 12 years of
school and 10% had 9 years or less of school education. Regarding
marital status, 32% of the sample was unmarried, 56% was mar-
ried, and 12% was divorced or widowed.
Women in the intervention group did not differ from those in the
control group in terms of BMI, age, education, or marital status.
2MATA ET AL.
tapraid5/zg1-hea/zg1-hea/zg100509/zg12352d09z
xppws S⫽1 7/8/09 22:47 Art: 2009-1642

APA PROOFS
There were also no differences between the 208 women who
completed the 12-month intervention and the 31 who quit the
program for any demographic or baseline psychosocial variable,
with the exception of age; women who stayed in the program were
on average 4 years older ( p⫽.01).
Measurements
Psychosocial Measures
General self-determination was assessed with the Self-
Determination Scale (Sheldon, Ryan, & Reis, 1996) which evalu-
ates individual differences for functioning in a self-determined
way. That is, being aware of one’s sense of self and feeling a sense
of choice toward one’s behaviors. For each of 10 statement pairs
such as “I always feel like I choose the things I do” and “I
sometimes feel that it’s not really me choosing the things I do”
participants evaluated each pair on a 5-point scale from “only A
feels true” to “only B feels true” (Cronbach’s alpha ⫽.63).
Reasons for staying in treatment (autonomous vs. controlled)
were measured using the Treatment Self-Regulation Questionnaire
(Williams, Freedman, & Deci, 1998; Williams et al., 1996) which
consists of 13 items and assesses the degree to which a person’s
motivation for participating in treatment is autonomous. On a
5-point scale, participants are asked to evaluate how well each
statement represents their reasons for staying in the program (e.g.,
“I would have felt bad about myself if I didn’t.”). The question-
naire consists of two subscales, autonomous (␣⫽.86) and con-
trolled (␣⫽.80) treatment self-regulation.
Exercise autonomous versus controlled self-regulation was
assessed with the Self-Regulation Questionnaire for Exercise
(adapted from Ryan & Connell, 1989). Items such as “I exer-
cise...because I simply enjoy working out” were evaluated on
a 7-point scale, ranging from not at all true to absolutely true.
The scale can be divided into two subscales, autonomous (Cron-
bach’s alpha ⫽.91) and controlled exercise self-regulation
(Cronbach’s alpha ⫽.72).
Exercise intrinsic motivation was measured with the Intrinsic
Motivation Inventory (Ryan, 1982; Ryan & Connell, 1989). The
questionnaire consists of 16 items measuring enjoyment, compe-
tence, involvement, and (absence of) pressure toward exercise, and
yielding an overall score of intrinsic motivation, used in this study.
It includes items such as “I think I’m good at being physically
active compared to other people,” evaluated on a 5-point scale
from not totally agree to totally disagree (␣⫽.94).
Eating behavior was measured with the Three-Factor Eating
Questionnaire (Stunkard & Messick, 1985). The 51-item scale is
divided into three subscales: cognitive restraint (␣⫽.82), cogni-
tive disinhibiton (␣⫽.68), and perception of hunger (␣⫽.78).
Statements include “On social occasions, like parties, I generally
eat too much,” that are evaluated on a 4-point scale from agree to
disagree or “If I ate too much on one day I try to make up for it on
the next day,” with answer format “true” or “false.” Because in this
study we were primarily interested in measuring markers of the
cognitive control of eating behavior, we did not use the scale
perception of hunger, which is highly influenced by physiological
states.
The Dutch Eating Behavior Questionnaire (Van Strien, Frijters,
Bergers, & Defares, 1986) was applied to assess external eating
(Cronbach’s alpha ⫽.88) and emotional eating (␣⫽.95). It
consists of 31 questions such as “Do you have a desire to eat when
you are irritated?” Answers are given on a 5-point scale from
“never” to “very frequently.”
Eating self-efficacy, the belief in one’s capacity for changing
eating behavior, was assessed with the Weight Management Effi-
cacy Questionnaire (Clark, Abrams, Niaura, Eaton, & Rossi,
1991). Statements include “I can resist food when I’m nervous,” to
be evaluated on a 10-point scale from not at all confident to very
confident. A global score including all items was used (␣⫽.95).
Exercise/Physical Activity
Minutes per week of leisure-time moderate and vigorous phys-
ical activities were estimated with the 7-Day Physical Activity
Recall interview (Blair et al., 1998; A. Dunn et al., 1999). Habitual
activities with a MET value above 3.0 and performed during the
last 7 days (or on a typical week of the past month) were quantified
to produce this variable.
Statistical Analyses
Treatment motivation was included in the analyses to test for the
effects of exercise motivation on eating self-regulation indepen-
dent of more autonomous versus controlled reasons to remain in
the program, thus making the test of our hypotheses more conser-
vative. Given that exercise was an integral part of the treatment in
this weight loss program, there might have been some overlap
between treatment regulations (more internal or more externally
controlled reasons to participate in the program) and exercise
regulations. The present analyses help distinguish these processes.
Twelve-month scores were used for all analyses. This choice
was based on the fact that not all psychosocial variables were
assessed at baseline. Most participants did not engage in regular
exercise at the beginning of the intervention, which yielded exer-
cise self-regulation measures less valid (e.g., “I exercise because
I” . . .). Also, treatment self-regulation (i.e., reasons to stay in
treatment) could only be assessed after the start of the intervention.
For consistency, we decided to also use physical activity measures
at 12 months, instead of change in physical activity. Because this
sample was mostly sedentary at baseline, the outcome (12-month)
measure was considered to represent well the result of the inter-
vention for this variable.
Stepwise hierarchical linear regressions with general self-
determination, treatment motivation, and exercise autonomous
motivation were used to test our first hypothesis. Because of
covariance in predictors (see Table 1), they were entered into the
model in a stepwise fashion; for models in which general self-
determination, treatment and exercise specific variables were en-
tered as predictors, general self-determination and treatment mea-
sures were entered first and exercise predictors in a second step.
This was done to determine the explanatory power of exercise
regulation variables above and beyond general self-determination
and treatment motivation.
Analyses were conducted for all participants (intervention and
controls) together. This was done to preserve statistical power and
increase variability in all measures under analysis, and also be-
cause the associations under scrutiny (self-determination as a pre-
dictor of eating self-regulation) were hypothesized to hold constant
3
EXERCISE SELF-DETERMINATION AND EATING SELF-REGULATION
AQ: 2
AQ: 3
T1, AQ:4
tapraid5/zg1-hea/zg1-hea/zg100509/zg12352d09z
xppws S⫽1 7/8/09 22:47 Art: 2009-1642

APA PROOFS
regardless of group membership. Still, it is possible that the biva-
riate relationships between general self-determination, treatment
and exercise specific motivation, and eating variables were con-
founded by the intervention effect, which could have influenced
both. Thus, regression analyses were further adjusted by group
membership.
To examine whether general self-determination, treatment, and
exercise-specific motivation mediated the relationship between
physical activity and eating self-regulation, multiple mediation
with tests of indirect effects were conducted (Preacher & Hayes,
2008). This procedure tests the first two formal steps of mediation
(predictor to mediator, mediator to outcome) and then provides
total, direct (not mediated) and indirect effects of the predictor
(physical activity) on outcomes (eating variables). The latter ef-
fects are then tested for significance, providing a formal test of
indirect or mediated effects of the predictor on the outcomes.
Results
In the intervention group, 86% of participants attended more
than 75% of the intervention sessions. In the control group (for
whom attending health education sessions was not mandatory),
20% of participants attended more than 75% of the sessions. At 12
months (end of intervention), the intervention group had increased
weight loss (⫺7.3% of initial body weight) and higher levels of
physical activity/exercise (M⫽300 min/wk moderate plus vigor-
ous exercise; M⫽9932 steps/day) than participants in the control
group (⫺1.7% of initial body weight; M⫽162 min/wk moderate
plus vigorous exercise; M⫽7852 steps/day; all ps⬍0.001).
Group differences in main intervention targets were medium to
large favoring the intervention group (all ps⬍0.001), including
general self-determination (d⫽0.40), and autonomous self-
regulation for treatment (d⫽1.35) and exercise (d⫽1.08). We
also found group differences in eating-related variables (all ps⬍
.001); the intervention group had higher eating self-efficacy (d⫽
0.64), higher cognitive restraint (d⫽0.48), lower disinhibition
(d⫽⫺0.53), lower emotional eating (d⫽⫺0.28), and lower
external eating scores (d⫽⫺0.66). The effects of the intervention
trial are reported in detail elsewhere (Silva et al., 2009).
There were no baseline differences between intervention and
control group for all predictor and dependent variables used in this
study, except for exercise intrinsic motivation, t(205) ⫽⫺2.04,
p⫽.04. However, effect size was small (d⫽0.28) and there was
no baseline difference in autonomous exercise self-regulation;
therefore, this difference was not interpreted. Table 1 shows in-
tercorrelations among all variables in the study.
Table 2 shows the results for stepwise multiple regression models,
separately for three different models as predictors of eating self-
regulation: General and treatment self-determination (Model A),
exercise-specific self-determination (Model B), and general, treat-
ment, and exercise-specific self-determination (Model C). Measures
of general self-determination and treatment motivation consistently
predicted eating self-regulation variables with all relationships in the
expected direction: positive relationships between measures of auton-
omy and eating variables typically associated with successful weight
management (cognitive restraint and eating self-efficacy) and nega-
tive relationships for hindering eating variable (disinhibition, emo-
tional and external eating). With the exception of disinhibition and
restraint, eating measures were predicted by general self-
determination. Every eating measure was predicted by at least one
measure of treatment motivation (autonomous or controlled). Intrinsic
exercise motivation (or autonomous exercise self-regulation) also
predicted all eating variables; however, the percent variance ac-
counted for by exercise-specific measures was slightly lower than that
observed for general and treatment-related measures. Nevertheless,
the exercise-specific measures which entered the model (i.e., exercise
intrinsic motivation) generally predicted eating self-regulation even
after accounting for general self-determination and treatment motiva-
tion.
To test whether these relationships hold when adjusting for
group membership (i.e., controlling for the intervention effect),
Table 1
Correlation Matrix of the Variables in the Study
2 3 4 5 6 7 8 9 10 11 12 13
1. General self-determination .30
ⴱⴱ
⫺.17
ⴱ
.24
ⴱⴱ
⫺.11 .31
ⴱⴱ
.35
ⴱⴱ
.25
ⴱⴱ
⫺.26
ⴱⴱ
⫺.25
ⴱⴱ
⫺.23
ⴱⴱ
.17
ⴱ
⫺.20
ⴱⴱ
2. Autonomous treatment self-
regulation — .10 .42
ⴱⴱ
.15
ⴱ
.59
ⴱⴱ
.37
ⴱⴱ
.35
ⴱⴱ
⫺.37
ⴱⴱ
⫺.29
ⴱⴱ
⫺.18
ⴱ
.35
ⴱⴱ
⫺.43
ⴱⴱ
3. Controlled treatment
self-regulation — ⫺.04 .57
ⴱⴱ
.04 ⫺.17
ⴱ
⫺.07 .15
⫹
.14
⫹
.19
ⴱⴱ
⫺.11 .02
4. Intrinsic motivation inventory — .07 .72
ⴱⴱ
.36
ⴱⴱ
.29
ⴱⴱ
⫺.28
ⴱⴱ
⫺.30
ⴱⴱ
⫺.24
ⴱⴱ
.41
ⴱⴱ
⫺.26
ⴱⴱ
5. Controlled exercise self-
regulation questionnaire — .19
ⴱⴱ
⫺.08 .11 .14
⫹
.08 .08 .03 .03
6. Autonomous exercise
self-regulation — .35
ⴱⴱ
.30
ⴱⴱ
⫺.29
ⴱⴱ
⫺.28
ⴱⴱ
⫺.17
ⴱ
.36
ⴱⴱ
⫺.32
ⴱⴱ
7. Eating self-efficacy — .41
ⴱⴱ
⫺.70
ⴱⴱ
⫺.68
ⴱⴱ
⫺.66
ⴱⴱ
.26
ⴱⴱ
⫺.37
ⴱⴱ
8. TFEQ-restraint — ⫺.25
ⴱⴱ
⫺.43
ⴱⴱ
⫺.19
ⴱⴱ
.30
ⴱⴱ
⫺.36
ⴱⴱ
9. TFEQ-disinhibition — .64
ⴱⴱ
.66
ⴱⴱ
⫺.19
ⴱ
.28
ⴱⴱ
10. DEBQ-external eating — .57
ⴱⴱ
⫺.27
ⴱⴱ
.22
ⴱⴱ
11. DEBQ-emotional eating — ⫺.15
ⴱ
.24
ⴱⴱ
12. Minutes of physical activity — ⫺.36
ⴱⴱ
13. Weight change (%, 0–12 mo.)
Note. For weight change (%, 0 –12 months): negative numbers represent weight loss, positive numbers weight gain.
⫹
p⬍.10 level (two-tailed).
ⴱ
p⬍.05 (two-tailed).
ⴱⴱ
p⬍.01.
4MATA ET AL.
T2
tapraid5/zg1-hea/zg1-hea/zg100509/zg12352d09z
xppws S⫽1 7/8/09 22:47 Art: 2009-1642

APA PROOFS
the same regression models were run, this time with the group
membership always forced into the model as one predictor.
Results were comparable to the unadjusted models; in particu-
lar, all five eating regulation measures that were predicted by
intrinsic motivation for exercise without group adjustment were
still predicted by exercise motivation when controlling for
group (all ␤s between .16 and .21 or ⫺.16 and ⫺.21, all
ps⬍.05; results not shown).
To test our second hypothesis, that is, whether general,
treatment, and exercise-specific measures mediated the rela-
tionship between physical activity and eating self-regulation,
mediation analyses were conducted. Variables that most con-
sistently predicted eating self-regulation in the regression anal-
yses were chosen as putative mediators (see Table 3). Physical
activity (the predictor) was significantly correlated to all me-
diators, general self-determination, autonomous treatment reg-
ulation, and intrinsic exercise motivation (A). All mediators
were significantly related to all eating regulation variables (the
outcomes), except intrinsic exercise motivation was not related
to restraint and autonomous treatment regulation was not asso-
ciated with emotional eating (B). As hypothesized, physical
activity was associated with all eating regulation variables in
the expected directions (C). When the mediators were added to
the model, the relationship between physical activity and all
types of eating regulations became nonsignificant (D). Results
for indirect effects (i.e., the magnitude of the mediation effect)
for individual mediators were significant for most of the eating
regulation variables (E). General self-determination was not
significant for disinhibition and external eating, and autonomous
treatment motivation not significant for emotional eating. Collec-
tively, results showed that overall general self-determination, auton-
omous treatment motivation, and intrinsic exercise motivation
fully mediated the relationship between physical activity and eat-
ing regulation.
Discussion
As hypothesized, exercise motivation and self-regulation was
associated with several important markers of eating self-
regulation. Exercise intrinsic motivation predicts eating regulation
beyond general self-determination and treatment motivation. This
suggests that not only general self-determination and treatment
Table 2
Stepwise Regression Analyses. Self-Determined Motivation as Predictor for Eating Self-Regulation
Dependent variable Predictors
Model A
General treatment
& self-determination
Model B
Exercise-specific
self-determination
Model C
General treatment
& exercise-specific
self-determination
␤p␤p␤P
Eating self-efficacy Self-determination (SDS) 0.23 .001 not tested 0.23 .001
Autonomous treatment self-regulation (TSRQ) 0.31 ⬍.001 not tested 0.24 .002
Controlled treatment self-regulation (TSRQ) ⫺0.15 .02 not tested not entered
Autonomous exercise self-regulation (ExSRQ) not tested 0.38 ⬍.001 not entered
Intrinsic motivation (IMI) not tested not entered 0.19 .01
R
2
(p) .23 (⬍.001) .15 (⬍.001) .27 (⬍.001)
TFEQ-restraint Self-determination (SDS) not entered not tested 0.12 .10
Autonomous treatment self-regulation (TSRQ) 0.34 ⬍.001 not tested 0.24 .003
Intrinsic motivation (IMI) not tested 0.29 ⬍.001 0.16 .004
R
2
(p) .12 (⬍.001) .10 (⬍.001) .17 (⬍.001)
TFEQ-disinhibition Autonomous treatment self-regulation (TSRQ) ⫺0.38 ⬍.001 not tested ⫺0.32 ⬍.001
Controlled treatment self-regulation (TSRQ) 0.17 .01 not tested 0.14 .05
Autonomous exercise self-regulation (ExSRQ) not entered ⫺0.33 ⬍.001 not entered
Controlled exercise self-regulation (ExSRQ) not tested 0.17 .02 not entered
Intrinsic motivation (IMI) not tested not entered ⫺0.16 .04
R
2
(p) .16 (⬍.001) .12 (⬍.001) .16 (⬍.001)
DEBQ-external eating Self-determination (SDS) ⫺0.24 .001 not tested ⫺0.14 .06
Autonomous treatment self-regulation (TSRQ) ⫺0.18 .01 not tested ⫺0.19 .01
Intrinsic motivation (IMI) not tested ⫺0.30 ⬍.001 ⫺0.19 .01
R
2
(p) .12 (⬍.001) .09 (⬍.001) .15 (⬍.001)
DEBQ-emotional eating Self-determination (SDS) ⫺0.15 .04 not tested ⫺0.15 .05
Autonomous treatment self-regulation (TSRQ) ⫺0.16 .03 not tested ⫺0.09 .23
Controlled treatment self-regulation (TSRQ) 0.18 .01 not tested 0.17 .02
Intrinsic motivation (IMI) not tested ⫺0.26 ⬍.001 ⫺0.19 .01
R
2
(p) .10 (⬍.001) .06 (⬍.001) .12 (⬍.001)
Note. Model A: General and treatment self-determination as predictors of eating regulation. Model B: Exercise-specific self-determination as predictor
of eating regulation. Model C: General, treatment, and exercise-specific self-determination as predictor of eating regulation. SDS, TSRQ autonomous and
controlled were always tested as predictors of eating regulation for Model A and C; IMI, ExSRQ autonomous and controlled, were always tested as
predictors of eating regulation for Model B and C. If a predictor is not listed, it did not enter any of the three models. “not-tested” means that this variable
was not tested in the specific regression model. “not entered” means that this variable was tested in this specific stepwise regression model but did not enter
the model due to poor empirical fit.
5
EXERCISE SELF-DETERMINATION AND EATING SELF-REGULATION
T3
tapraid5/zg1-hea/zg1-hea/zg100509/zg12352d09z
xppws S⫽1 7/8/09 22:47 Art: 2009-1642

APA PROOFS
Table 3
Summary of Mediation Analyses
A. Predictor (min moderate and vigorous physical activity) to mediators (motivation)
␤SE
General self-determination .19
ⴱ
.08
Autonomous treatment regulation .34
ⴱⴱ
.07
Intrinsic motivation to exercise .40
ⴱⴱ
.06
B. Direct effects of mediators (self-determined motivation) to outcomes (eating variables)
Eating self-efficacy TFEQ-restraint TFEQ-disinhibition DEBQ-external eating DEBQ-emotional eating
␤SE ␤SE ␤SE ␤SE ␤SE
General self-determination .22
ⴱⴱ
.07 .12
⫹
.07 ⫺.13
⫹
.08 ⫺.14
⫹
.07 ⫺.18
ⴱ
.07
Autonomous treatment regulation .24
ⴱⴱ
.08 .21
ⴱⴱ
.08 ⫺.29
ⴱⴱ
.08 ⫺.17
ⴱ
.08 ⫺.07 .08
Intrinsic motivation exercise .17
ⴱ
.08 .11 .08 ⫺.15
⫹
.09 ⫺.16
⫹
.09 ⫺.21
ⴱ
.09
C. Total effect of predictor (physical activity) on outcomes (eating variables)
Eating self-efficacy TFEQ-restraint TFEQ-disinhibition DEBQ-external eating DEBQ-emotional eating
␤SE ␤SE ␤SE ␤SE ␤SE
Min moderate and vigorous
physical activity .26
ⴱⴱ
.08 .27
ⴱⴱ
.07 ⫺.19
ⴱ
.08 ⫺.28
ⴱⴱ
.07 ⫺.15
ⴱ
.07
D. Direct effect of predictor (physical activity) on outcomes (eating variables)
Eating self-efficacy TFEQ-restraint TFEQ-disinhibition DEBQ-external eating DEBQ-emotional eating
␤SE ␤SE ␤SE ␤SE ␤SE
Min moderate and vigorous
physical activity .06 .08 .13
⫹
.07 ⫺.02 .08 ⫺.13 .08 ⫺.01 .08
E. Normal theory tests for indirect effects
Eating self-efficacy TFEQ-restraint TFEQ-disinhibition DEBQ-external eating DEBQ-emotional eating
␤SE ␤SE ␤SE ␤SE ␤SE
Total indirect effect .20
ⴱⴱ
.05 .14
ⴱⴱ
.04 ⫺.18
ⴱⴱ
.05 ⫺.15
ⴱⴱ
.04 ⫺.14
ⴱⴱ
.04
General self-determination .05
ⴱ
.02 .02
ⴱ
.02 ⫺.02 .02 ⫺.02 .02 ⫺.03
⫹
.02
Autonomous treatment regulation .08
ⴱ
.03 .07
ⴱ
.03 ⫺.10
ⴱⴱ
.04 ⫺.06
⫹
.03 ⫺.02 .03
Intrinsic motivation exercise .08
ⴱ
.04 .04
ⴱ
.03 ⫺.06
⫹
.03 ⫺.06
⫹
.04 ⫺.08
ⴱ
.04
Note.
⫹
p⬍0.10 (all two-tailed).
ⴱ
p⬍0.05 (two-tailed).
ⴱⴱ
p⬍0.01.
6MATA ET AL.
tapraid5/zg1-hea/zg1-hea/zg100509/zg12352d09z
xppws S⫽1 7/8/09 22:47 Art: 2009-1642

APA PROOFS
motivation “spilled-over” to eating regulation, but also that
exercise-specific motivation additionally contributed to improved
eating behavior. Furthermore, the relationship between self-
reported physical activity and eating regulation is mediated by
general self-determination, autonomous treatment motivation, and
(for most eating-related outcomes) intrinsic exercise motivation.
This suggests that, besides physiological effects of exercise which
may affect appetite regulation, motivational mechanisms may also
explain the association between physical activity and eating be-
haviors.
When controlling for intervention effects, the unique contribution
of exercise motivation for eating regulation persists, suggesting
that increase in exercise motivation is associated with eating
regulation independent of the intervention treatment. One inter-
esting difference in the group-adjusted analyses was that autono-
mous motivation for treatment was no longer predictive of emo-
tional eating and external eating. This suggests a strong
intervention effect on these particular variables, a fact consistent
with the intervention curriculum which covered these topics to a
considerable extent and in various sessions (Silva et al., 2008).
Our study’s results relate to findings from other weight manage-
ment intervention trials suggesting that changing both eating and
exercise behavior might have synergistic effects (C. L. Dunn et al.,
2006; Jakicic et al., 2002) or longer-lasting effects than diet or
exercise change alone (see Miller, Koceja, & Hamilton, 1997, for a
meta-analysis). However, few studies have included or reported on
the effect of psychological factors, such as motivation, on multiple
behavior change (J. O. Prochaska et al., 2008), thus identifying
potential mechanisms underlying this synergistic change. Taking mul-
tiple behavior intervention from a behavioral to a motivational level is
a promising step to discover underlying mechanisms, which in turn
can be powerful targets for more successful, long-term interventions.
Qualitative changes in motivation, from less to more autonomous,
may not happen in isolation but instead apply to various domains
simultaneously, even if to varying degrees. In this study, more auton-
omous general self-determination, treatment motivation, and exercise-
specific intrinsic motivation resulting from a theory-guided interven-
tion were related to several eating regulation variables. The predictive
power of general self-determination for eating regulation could sup-
port Vallerand’s (1997) assumption of motivational bottom-up pro-
cesses (e.g., an increase in exercise specific self-determination affect-
ing increased general self-determination), which would in turn
influence eating regulation (through top-down processes). However,
exercise-specific motivation was associated with eating regulation
beyond change in general self-determination (although explaining
fewer variance than general self-determination), thus suggesting also
a dynamic interplay between the two contextual levels (i.e., physical
activity and eating) of motivation, in line with Hagger and colleagues’
trans-contextual model (Hagger et al., 2003).
General and specific motivational change resulting from interven-
tions is one plausible mechanism underlying the association between
increased physical activity and improved eating regulation. In future
studies, it could be promising to measure psychological factors simul-
taneously with other factors that may be involved in the relationship
between physical activity and eating regulation, such as physiological
(e.g., appetite regulation) or behavioral (e.g., stress management strat-
egies), to compare their relative influences on eating behavior. It
should be noted that although we specifically wanted to explore the
hypothesis that led from exercise behavior and exercise motivation to
eating behavior, one cannot exclude reciprocal effects at the motiva-
tional level, where, for instance, success at eating self-regulation
would also positively influence motivation and/or confidence for
exercising.
Our results concerning self-reported physical activity echo previous
observational research suggesting physical activity as a gateway behavior
for motivational changes in eating regulation (Blakely, Dunnagan,
Haynes, Moore, & Pelican, 2004; Nigg et al., 1999; Tucker & Reicks,
2002). However, intervention studies have not generally found physical
activity to have such a gateway function (Dutton, Napolitano, Whiteley,
& Marcus, 2008; Wilcox, King, Castro, & Bortz, 2000). One possible
explanation for these inconsistent findings is that motivational change
toward health behaviors was not the target of the intervention trials but
rather implementation of behavioral programs. For example, Wilcox and
colleagues (2005) reported that both physical activity and eating behavior
changed after a behavioral intervention targeting physical activity. How-
ever, change in physical activity did not explain change in eating behav-
ior, suggesting the existence of a third factor that would underlie the
change in eating through physical activity change. Such a factor could be
motivational in nature, namely the extent to which motivation regulation
is more self-directed and less contingent on external demands. A strictly
behavior-focused intervention, not directed at creating an autonomy-
promoting climate may insufficiently influence internal self-regulation
and intrinsic motivation.
In conclusion, this study shows that spill-over effects may occur
between treatment and exercise motivation and eating self-regulation, in
the course of a weight control intervention. Furthermore, the qualitative
nature of motivational regulation (i.e., intrinsic and autonomous vs. ex-
ternally controlled) seem to be underlying mechanisms for the relation-
ship between actual physical activity and eating regulation. Research
concerning multiple behavior change is the future of preventive medicine
(J. O. Prochaska, 2008). However, so far there has been little effort to
develop theories of health behavior that directly address intervention in
more than one behavior simultaneously (J. J. Prochaska, Spring, & Nigg,
2008) or programmatic research showing the effectiveness of interven-
tions targeting two or more health behaviors (J. O. Prochaska, 2008).
Investigating motivation in more detail in the context of behavioral
weight management programs holds promise for the development of
psychological models of multiple behavior change. It also has direct
applied value, informing possible effective strategies for weight manage-
ment in clinical practice.
References
Annesi, J. J., & Unruh, J. L. (2008). Relations of exercise, self-appraisal,
mood changes and weight loss in obese women: Testing propositions
based on Baker and Brownell’s 2000 model. The American Journal of
Medical Sciences, 335, 198 –204.
Baker, C. W., & Brownell, K. D. (2000). Physical activity and maintenance
of weight loss: Physiological and psychological mechanisms. In B.
Christopher (Ed.), Physical Activity and Obesity (pp. 311–328). Cham-
paign, IL: Human Kinetics.
Blair, S., Applegate, W., Dunn, A., Ettinger, W., Haskell, W., King, A.,
et al. (1998). Activity Counseling Trial (ACT): Rationale, design, and
methods. Medicine & Science in Sports and Exercise, 30, 1097–1106.
Blakely, F., Dunnagan, T., Haynes, G., Moore, S., & Pelican, S. (2004).
Moderate physical activity and its relationship to select measures of a
healthy diet. Journal of Rural Health, 20, 160 –165.
Blanchard, C., Mask, L., Vallerand, R. J., de la Sablionnie`re, R., &
Provencher, P. (2007). Reciprocal relationships between contextual and
7
EXERCISE SELF-DETERMINATION AND EATING SELF-REGULATION
AQ: 5
tapraid5/zg1-hea/zg1-hea/zg100509/zg12352d09z
xppws S⫽1 7/8/09 22:47 Art: 2009-1642

APA PROOFS
situational motivation in a sports setting. Psychology of Sport & Exer-
cise, 8, 854 – 873.
Clark, M. M., Abrams, D. B., Niaura, R. S., Eaton, C. A., & Rossi, J. S.
(1991). Self-efficacy in weight management. Journal of Consulting and
Clinical Psychology, 59, 739 –744.
Deci, E., & Ryan, R. (1985). Intrinsic motivation and self-determination in
human behavior. New York: Plenum Press.
Dunn, A., Marcus, B., Kampert, J., Garcia, M., Kohl, H., & Blair, S.
(1999). Comparison of lifestyle and structured interventions to increase
physical activity and cardiorespiratory fitness. Journal of the American
Medical Association, 281, 327–334.
Dunn, C. L., Hannan, P. J., Jeffery, R. W., Sherwood, N. E., Pronk, N. P.,
& Boyle, R. (2006). The comparative and cumulative effects of a dietary
restriction and exercise on weight loss. International Journal of Obesity,
30, 112–121.
Dutton, G. R., Napolitano, M. A., Whiteley, J. A., & Marcus, B. H. (2008).
Is physical activity a gateway behavior for diet? Findings from a phys-
ical activity trial. Preventive Medicine, 46, 216 –221.
Fortier, M. S., Sweet, S. N., O’Sullivan, T. L., & Williams, G. C. (2007).
A self-determination process model of physical activity adoption in the
context of a randomized controlled trial. Psychology of Sport and Ex-
ercise, 8, 741–757.
Hagger, M. S., Chatzisarantis, N. L., Culverhouse, T., & Biddle, S. J. H.
(2003). The process by which perceived autonomy support in physical
education promotes leisure-time activity intentions and behavior: A trans-
contextual model. Journal of Educational Psychology, 95, 784 –795.
Hagger, M. S., Chatzisarantis, N. L. D., Barkoukis, V., Wang, C. K. J., &
Baranowski, J. (2005). Perceived autonomy support in physical education
and leisure-time physical activity: A cross-cultural evaluation of the trans-
contextual model. Journal of Educational Psychology, 97, 376 –390.
Jakicic, J. M., Wing, R. R., & Winters-Hart, C. (2002). Relationship of
physical activity to eating behaviors and weight loss in women. Medi-
cine & Science in Sports & Exercise, 34, 1653–1659.
Martins, C., Morgan, L., & Truby, H. (2008). A review of the effects of
exercise on appetite regulation: An obesity perspective. International
Journal of Obesity, 32, 1337–1347.
Miller, W. C., Koceja, D. M., & Hamilton, E. J. (1997). A meta-analysis of
the past 25 years of weight loss research using diet, exercise or diet plus
exercise intervention. International Journal of Obesity and Related
Metabolic Disorders, 21, 941–947.
Nigg, C. R., Burbank, P. M., Padula, C., Dufresne, R., Rossi, J. S., Velicer,
W. F., et al. (1999). Stages of change across ten health risk behaviors for
older adults. Gerontologist, 39, 473– 482.
Pelletier, L. G., & Dion, S. C. (2007). An examination of general and
specific motivational mechanisms for the relation between body dissat-
isfaction and eating behaviors. Journal of Social and Clinical Psychol-
ogy, 26, 303–333.
Pelletier, L. G., Dion, S. C., Slovinec-D’Angelo, M., & Reid, R. (2004).
Why do you regulate what you eat? Relationships between forms of
regulation, eating behaviors, sustained dietary behavior change, and
psychological adjustment. Motivation and Emotion, 28, 245–277.
Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling
strategies for assessing and comparing indirect effects in multiple me-
diator models. Behavioral Research Methods, 40, 879 – 891.
Prochaska, J. J., Spring, B., & Nigg, C. R. (2008). Multiple health behavior
change research: An introduction and overview. Preventive Medicine,
46, 181–188.
Prochaska, J. O. (2008). Multiple Health Behavior Research represents the
future of preventive medicine. Preventive Medicine, 46, 281–285.
Prochaska, J. O., Butterworth, S., Redding, C. A., Burden, V., Perrin, N.,
Leo, M., et al. (2008). Initial efficacy of MI, TTM tailoring and HRI’s
with multiple behaviors for employee health promotion. Preventive
Medicine, 46, 226 –231.
Pronk, N. P., Anderson, L. H., Crain, A. L., Martinson, B. C., O’Connor,
P. J., Sherwood, N. E., et al. (2004). Meeting recommendations for
multiple healthy lifestyle factors. Prevalence, clustering, and predictors
among adolescent, adult, and senior health plan members. American
Journal of Preventive Medicine, 27, 25–33.
Ryan, R. M. (1982). Control and information in the intrapersonal sphere:
An extension of cognitive evaluation theory. Journal of Personality and
Social Psychology, 43, 450 – 461.
Ryan, R. M., & Connell, J. (1989). Perceived locus of causality and
internalization: Examining reasons for acting in two domains. Journal of
Personality and Social Psychology, 57, 749 –761.
Ryan, R. M., & Deci, E. L. (2000). Intrinsic and extrinsic motivations:
Classic definitions and new directions. Contemporary Educational Psy-
chology, 25, 54 – 67.
Silva, M. N., Markland, D. A., Minderico, C. S., Vieira, P. N., Castro,
M. M., Coutinho, S. R., et al. (2008). A randomized controlled trial to
evaluate Self-Determination Theory for exercise adherence and weight
control: Rationale and intervention description. BMC Public Health, 8,
234.
Silva, M. N., Vieira, P. N., Coutinho, S. R., Matos, M. G., Sardinha, L. B.,
& Teixeira, P. J. (2009). Using self-determination theory to promote
physical activity and weight control: A randomized controlled trial in
women. Manuscript submitted for publication.
Stunkard, A., & Messick, S. (1985). The three-factor eating questionnaire
to measure dietary restraint, disinhibition and hunger. Journal of Psy-
chosomatic Research, 29, 71– 83.
Teixeira, P. J., Going, S. B., Houtkooper, L. B., Cussler, E. C., Metcalfe,
L. L., Blew, R. M., et al. (2006). Exercise motivation, eating, and body
image variables as predictors of weight control. Medicine & Science in
Sports & Exercise, 38, 179 –188.
Tucker, M., & Reicks, M. (2002). Exercise as a gateway behavior for
healthful eating among older adults: An exploratory study. Journal of
Nutrition Education & Behavior, 34, S14 –S19.
Vallerand, R. J. (1997). Toward a hierarchical model of intrinsic and
extrinsic motivation. In M. P. Zanna (Ed.), Advances in experimental
social psychology (pp. 271–360). New York: Academic Press.
Vallerand, R. J., & Ratelle, C. F. (2002). Intrinsic and extrinsic motivation:
A hierarchical model. In E. L. Deci & R. M. Ryan (Eds.), The motivation
and self-determination of behavior: Theoretical and applied issues.
Rochester, NY: University of Rochester Press.
Van Strien, I., Frijters, J., Bergers, G., & Defares, P. (1986). The Dutch
Eating Behavior Questionnaire (DEBQ) for assessment of restrained,
emotional and external eating behavior. International Journal Eating
Disorders, 5, 295–315.
Wilcox, S., King, A. C., Castro, C., & Bortz, W. (2000). Do changes in
physical activity lead to dietary changes in middle and old age? Amer-
ican Journal of Preventive Medicine, 18, 276 –283.
Williams, G. C., Freedman, Z. R., & Deci, E. L. (1998). Supporting
autonomy to motivate patients with diabetes for glucose control. Dia-
betes Care, 21, 1644 –1651.
Williams, G. C., Grow, V. M., Freedman, Z. R., Ryan, R. M., & Deci, E. L.
(1996). Motivational predictors of weight loss and weight-loss mainte-
nance. Journal of Personality and Social Psychology, 70, 115–126.
8MATA ET AL.
tapraid5/zg1-hea/zg1-hea/zg100509/zg12352d09z
xppws S⫽1 7/8/09 22:47 Art: 2009-1642