Psychology of Sport and Exercise 10 (2009) 226–234

Contents lists available at ScienceDirect

Psychology of Sport and Exercise

journal homepage: www.elsevier.com/locate/psychsport

The role of emotion in bridging the intention–behaviour gap: The case

of sports participation
Changiz Mohiyeddini a, *, Regina Pauli a, Stephanie Bauer b
a
Whitelands College, Roehampton University, Holybourne Avenue, London SW15 4JD, London, UK
b
Center for Psychotherapy Research, University Hospital Heidelberg, Bergheimer Street 54, 69115 Heidelberg, Germany

a r t i c l e i n f o a b s t r a c t

Article history: Problem: Psychological models of sports participation frequently draw on the Theory of Planned
Received 30 August 2007
Behaviour (TPB) in order to predict health-behaviour-related intentions. While these models commonly
Received in revised form 19 August 2008
show high predictive power with respect to intention, they often fall short in the prediction of behaviour
Accepted 22 August 2008
Available online 13 September 2008 (the ‘intention–behaviour gap’). The present study contends that integrating emotional processes into
TPB can substantially improve the model’s predictive power over and above traditional cognitive
predictors.
Keywords:
Intention–behaviour gap Method: In a longitudinal study structural equation modelling was employed to test a traditional model
Emotions (based on TPB) and an extended model including emotion associated with the intention to exercise as
Theory of planned behaviour a mediator variable. A community sample (N ¼ 237) completed a questionnaire measuring the compo-
Physical exercise nents of a traditional TPB model. Prior to each of 16 expected exercise sessions (over 8 weeks) intention
Health behaviour to exercise and emotion associated with the intention were measured.
Maintenance of physical activity
Results: Results confirmed a substantial increase of 17% in explained variance of exercise frequency and
20% in exercise duration for the extended model. Emotional appraisal of the intention to exercise thus
appears to mediate the traditional intention–behaviour relationship, suggesting that emotionally based
interventions aimed at increasing sports participation may be helpful.
Conclusion: It is concluded that emotion variables should be added to traditional TPB models in order to
predict health behaviour more fully.
Ó 2008 Elsevier Ltd. All rights reserved.

Regular physical activity is an important component of a healthy
lifestyle (Pate, Dowda, O’Neill, & Ward, 2007). Empirical results
show that participation in physical activity is low in significant
proportions of the population (Dishman & Buckworth, 2001; Mar-
tinez-Gonzalez et al., 2001) and that interventions designed to
increase sports participation do not always have the desired effect
on behaviour despite increasing intention in participants to become
physically more active (Milne, Orbell, & Sheeran, 2002).
A number of social and health behaviours have been studied in
relation to the theory of planned behaviour (TPB, Ajzen, 1991),
which contends that intention is a primary predictor of subsequent
behaviour (Armitage & Conner, 2000; Wallston & Armstrong,
2002). TPB postulates that behaviour is determined by behavioural
intentions and perceived behavioural control (PBC). Intentions are
defined as explicit, behaviourally orientated decisions to act, i.e.
formulations of goal states which are achievable through behav-
iour. They are independently influenced by subjective norms,
* Corresponding author. Tel.: þ44 (0) 20 8392 3616; fax: þ44 (0) 20 8392 3527.
E-mail address: c.mohiyeddini@roehampton.ac.uk (C. Mohiyeddini).
which relate to perceived expectations of others, and attitudes
which can be positive or negative towards the specific intended
behaviour. PBC, which is defined as the perception of subjective
control at the point of translation of goal states into actual behav-
iour, predicts behaviour both directly and via intentions (Hagger,
Chatzisarantis, & Biddle, 2002).
Applications of TPB to the relationship between intention to
exercise and subsequent behaviour have revealed that 45% of the
variance in intention can be predicted from PBC, subjective norms
and attitudes but only 27% of the variance in exercise behaviour can
be predicted on the basis of intention (based on Hagger et al.’s
(2002) meta-analysis of 72 studies). Sutton’s (1998) review also
indicated that 19–38% of the variance in behaviour can be explained
by intention and PBC. Similarly, in a recent longitudinal study 49%
of the variance in intention could be explained by PBC, subjective
norms and attitudes, whereas only 22% of the variance in actual
exercise behaviour could be explained (Armitage, 2005). In
a related vein, an experimental study conducted by Chatzisarantis
and Hagger (2005) demonstrated that attitudes can predict inten-
tions sufficiently whereas intentions do not always lead to physical
activity.

1469-0292/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.psychsport.2008.08.005
 C. Mohiyeddini et al. / Psychology of Sport and Exercise 10 (2009) 226–234
The implication is that intention formation is well understood as
it can be well predicted, but the processes involving the translation
of intentions into behaviour are not fully accounted for by current
models. This ‘‘intention–behaviour gap’’ (Sheeran, 2002) is also
evident in a lack of temporal stability of the intention–behaviour
relationship. Recent research has focused on cognitive variables as
mediators of the intention–behaviour relationship, e.g. motiva-
tional and volitional variables. Milne et al. (2002) showed that
interventions directed at volitional (aimed at implementation
intention) as well as motivational variables significantly increased
subsequent likelihood of exercise behaviour compared with inter-
ventions to increase motivation alone. Moreover, empirical
evidence has emerged, showing that self-regulation processes
(Abraham, Sheeran, & Johnston, 1998) and perceived self-efficacy
(Coumeya & McAuley, 1994; Estabrooks & Carron, 1998) can explain
additional variance in behaviour.
Emotions and the intention–behaviour gap
It is generally agreed in psychology that behaviour can be
generated through multiple processes. Hence it is our contention
that emotions act as an important missing link in the intention–
behaviour gap. Conner and Armitage (1998) have already pointed
out, that an important shortcoming of TPB is its exclusion of
affective processes, which are crucial for the intentional process
(Van der Pligt, Zeelenberg, van Dijk, de Vries, & Richard, 1998). Even
though attitudes as a component of TPB are usually described to
have an affective component (Eagly & Chaiken, 1993), this
component seems to be so closely correlated to cognitive and
behavioural components of attitudes that its (unique) impact on
behaviour is hardly measurable (e.g. Dillon & Kumar, 1985). Hence
we suggest adding emotions as a separate component to the TPB.
In line with Perugini and Bagozzi (2001), the nature and role of
emotions as suggested here differ from attitudes mainly in two
respects: 1) while our specification of emotions is focused on the
intention to exercise, attitudes focus directly on the actual behav-
iour, 2) emotions are considered as short-term and dynamic
processes. In contrast, attitudes are assumed to be learnt and
disposition-like reactions to specific objects or actions. A further
difference between emotions and attitudes in the present study
pertains to the measurement of these constructs. In accordance
with Perugini and Bagozzi (2001), we suggest using uniploar items
to assess emotions (for example, the experience of e.g. ‘alertness’
along a continuum from ‘not at all’ to ‘extremely’) and bipolar items
to assess attitudes.
The questions on how cognition (e.g. intention) and emotion
interrelate and on the function of emotion in relation to action are
currently unresolved and controversial questions in theories of
emotion (Leventhal & Scherer, 1987; Scherer, Schorr, & Johnstone,
2001; Solomon, 2000). However, the way in which intentions may
cause an emotional reaction which in turn affects subsequent
behaviour is addressed in Bagozzi’s theory of self-regulation
(Bagozzi, 1992; Bagozzi, Baumgartner, & Pierters, 1998; Bagozzi,
Moore, & Leone, 2004). This theory suggests that behaviour can be
conceived as a goal-directed activity and can be studied from the
perspective of trying to act, where people set goals and make plans
to implement their decisions. According to this theory, the inten-
tion–emotion link may depend on the success (or failure) of the
implementation of these decisions, i.e. ultimately on the success (or
failure) of goal attainment. This success requires planning and
continuous commitment despite potential (internal and external)
barriers as well as the self-managed ability to resist upcoming
temptations (Bagozzi et al., 2004).
Consequently, various impediments (e.g. lack of personal
competencies, situational barriers, contradicting intentions) can
lead to a negative emotional response such as frustration or
227
sadness. In contrast, positive emotional responses such as excite-
ment or happiness will appear in situations in which no serious
barriers counteract the implementation of a person’s intention.
Such positive responses can enhance a person’s motivation to act
(Chen & Bargh, 1999) e.g. by stimulating approach or avoidance
tendencies (see Carver, Sutton, & Scheier, 2000). Partly based on
Carver et al.’s (2000) work, we assume that the emotional states
that are activated by the formulation of intentions may support
initiating (and maintaining) exercise behaviour because one feels it
is a positive thing to do (approach tendency) or because of a need to
avoid the negative consequences of not doing so (avoidance
tendency). In line with this, Bagozzi et al. (1998) assume that the
attainment of a sub goal leads to positive emotions and to the
continuation of a person’s current plan, while the failure to achieve
a sub goal is associated with negative emotions. Consequently,
positive emotions may facilitate the translation of intentions to
exercise into action as well as increase the frequency and duration
of physical activity. In contrast, negative emotions may lead to
a decrease of the likelihood of commencing an exercise regime, as
well as to a decrease of the frequency of exercise sessions and the
duration of each exercise event.
There is some support for the assumption that emotions affect
intention and action. Zhu and Thagard (2002) have argued that
emotions significantly affect action generation (intention to act) as
well as action execution and control. This means that emotions are
considered to be positive or negative responses towards an inten-
ded action relevant in mediating the association between intention
and behaviour. Additional support stems from qualitative studies
concluding that emotional mediation of behaviour may play a role
in the maintenance of exercise behaviour. Specifically, positive
emotional outcomes were shown to be related to intention, moti-
vation and behaviour in participants engaging in outdoor aerobic
exercise (Marttila & Nupponen, 2000).
In sum, a variety of theoretical perspectives provide the basis for
assuming that the inclusion of emotion as a mediator between
intention and behaviour can lead to a significant increase in the
variance accounted for in exercise behaviour. Mohiyeddini and
Bauer (2007) have shown that including a single assessment of the
emotional states that are activated by the formulation of the
intention to exercise yielded a significant increase in the variance
explained compared with traditional TBP models based only on
cognitive predictors. This was the case despite a considerable time
lag (4 months) between assessment of predictors of intention,
intention and associated emotion on the one hand and measure-
ment of subsequent exercise duration on the other, indicating that
this may be a promising approach in addressing the intention–
behaviour gap with respect to exercise behaviour. However,
Mohiyeddini and Bauer (2007) excluded non-intenders and those
who failed to translate their intentions to exercise into actual
behaviour thus limiting the generalisability of their findings. They
suggest that a longitudinal study in which cognitive and emotional
components of the model are repeatedly assessed may serve to
resolve the issue of whether emotion is an important additional
variable in explaining variance in exercise frequency and duration.
Based on these considerations, we conducted a longitudinal study
to examine the predictive validity of the traditional TPB model as
well as the extended TPB model in which emotions were included
as a mediator variable.
The present research
In the present study the frequency and the duration of physical
activity, the intention to exercise as well as the emotional appraisal
of the intention were measured on 16 occasions (twice weekly)
over a two month period. In line with the TPB it was expected that
228 C. Mohiyeddini et al. / Psychology of Sport and Exercise 10 (2009) 226–234
intention to become physically active is influenced by attitudes,
subjective norms and PBC.
Furthermore we assumed that the manifest exercise behaviour
measured in terms of exercise frequency (as indicator for mainte-
nance of physical activity) and duration can be predicted from
intention to become physically active and PBC. It was further
hypothesised that emotions associated with the intention to
become physically active, serve as mediators (see Baron & Kenny,
1986) in the intention–behaviour relationship and their inclusion
leads to a substantial increase in the explained variance of exercise
frequency and duration.
Method
Participants
A German speaking Zurich community sample of 285 people
initially agreed to participate in this study in response to infor-
mation distributed via mailing lists. However 29 of these did not
respond to the initial researcher correspondence (see procedure). A
further 19 participants completed an initial questionnaire pack but
either failed to confirm their password electronically (n ¼ 14) or
made no further contact subsequently (n ¼ 5). No data is available
from these participants and they were therefore excluded from the
analysis. The remaining 237 participants were on average 26.94
years old (SD ¼ 9.9; range ¼ 17–69). Male participants (n ¼ 108)
yielded a mean of 26.25 years (SD ¼ 9.29; range ¼ 18–64) and the
mean age of the female participants (n ¼ 129) was 27.51 years
(SD ¼ 10.55; range ¼ 17–69). There were no age differences
between male and female participants (t(235) ¼ 0.97, n.s.).
Variables
Age and gender
Previous findings indicate that sample characteristics such as
age and gender may influence the relative contributions of the TPB
cognitions on explanation of intention and behaviour (Ajzen &
Fishbein, 1980; Henderson, Bialeschki, Shaw, & Freysinger, 1999;
Sheeran & Orbell, 1998). Additionally, it was found that engagement
in physical activity declines with age (Armstrong & McManus,
1994). In this study gender differences were explored by ANCOVA
with age as a covariate. The covariate was not significant (Fs(1,
234) < 3.53, n.s.). Furthermore, males and females did not differ in
attitudes, subjective norms, PBC, intention, emotions, training
duration or frequency (Fs(1, 234) < 3.18, n.s.). Therefore, gender and
age were excluded from further analysis.
Socioeconomic status (SES)
Previous studies have shown that physical activity is related to
SES. Low SES, i.e. low education level, income and/or blue-collar
occupation, was found to decrease the probability of physical
activity (Brown & Siahpush, 2006). Therefore we measured the level
of education, income and occupation. In the present study, 46.4% of
participants were employed or self-employed, 21.5% were students,
21.5% described themselves as home makers, 8% were unemployed
and 2.5% were retired. In terms of educational attainment, 33.3% of
participants held a university degree, 31.6% were educated to
tertiary level (Matura) and 35% had completed high school (a
minimum of 9 years of full time schooling).
For the assessment of income, we used a 6 point income scale.
The categories were chosen by the following portions of partici-
pants: up to 25000CHF (approx. USD 20000) per year: 1.7%, up to
40000CHF (approx. USD 33000): 17.3%, up to 55000CHF (approx.
USD 45000): 30.4%, up to 70000CHF (approx. USD 58000): 35.4%,
up to 85000CHF (approx. USD 70000): 12.7% and above 85000CHF:
2.5%. There were no differences between the occupational status
(Fs(5, 231) < 1.96, n.s.), educational attainment groups (Fs(2,
234) < 2.69, n.s.), and income groups (Fs(5, 231) < 2.31, n.s.) with
respect to the variables measured in this study and they were not
considered in any further analysis.
Previous experience with exercise
Participants were asked to evaluate their experience with
exercise on a four point scale (inexperienced, some experience,
experienced, highly experienced) due to findings that previous
behaviour significantly influences intentions, attitudes, subjective
norms, PBC, and behaviour (Bagozzi, 1981; Hagger, Chatzisarantis,
Biddle, & Orbell, 2001; Hagger et al., 2002). 36.3% of participants
described themselves as inexperienced, 40.1% as having some
experience, 16.5% as experienced and 7.2% as highly experienced.
The four groups did not differ in respect of any of the variables
measured in this study (Fs(3, 233) < 1.39, n.s.). Therefore previous
experience with exercise was not analyzed any further.
Attitudes, subjective norms and PBC
Attitudes, subjective norm and PBC were measured with
a slightly modified German version of Armitage’s items (see
Armitage, 2005, p. 237): Attitudes were measured using a 6 item
semantic differential scale. Participants rated adjective pairs in the
context of the following item anchor; ‘‘Exercising twice a week is
..... for me.’’. Adjective pairs included dull–interesting (item
A1), unpleasant–pleasant (item A2), boring–stimulating (item A3),
unhealthy–healthy (item A4), bad–good (item A5), and useless–
useful (item A6). These were rated on a seven point scale from 3 to
þ3. Cronbach’s alpha for the attitude scale was 0.74.
Subjective norms were measured with three items. These were
‘‘People close to me think I should participate in regular physical
activity’’ (disagree/agree) (item SN1); ‘‘People who are important to
me would.’’ (disapprove of my participating in regular physical
activity/approve of my participating in regular physical activity) (item
SN2); and ‘‘People close to me think I.’’ (should not participate in
regular physical activity /should participate in regular physical activity)
(item SN3). Subjective norm items were rated on a unipolar scale (þ1
to þ7). Cronbach’s alpha for the subjective norms scale was 0.84.
Four items were used to measure PBC. These were ‘‘To what
extent do you see yourself as being capable of participating in
regular physical activity?’’ (incapable–capable) (item PBC1); ‘‘How
confident are you that you will be able to participate in regular
physical activity?’’ (not very confident–very confident) (item PBC2);
‘‘I believe I have the ability to participate in regular physical
activity’’ (definitely do not–definitely do) (item PBC3); and ‘‘How
much personal control do you feel you have over participating in
regular physical activity?’’ (no control–complete control) (item
PBC4). All items were measured on unipolar (þ1 to þ7) scales.
Cronbach’s alpha for the PBC scale was 0.83.
Intention
Intention to engage in exercise was repeatedly assessed at 16
time points, prior to each expected exercise event. Based on Blan-
chard, Rodgers, Courneya, Daub, and Knapik (2002), intention was
measured with two items: ‘‘My goal is to exercise tomorrow for at
least 30 minutes’’ and ‘‘I intend to exercise tomorrow for at least 30
minutes’’. Participants rated both items on a seven point bipolar
scale ranging from 3 (not at all true for me) to þ3 (completely true
for me).
Emotional appraisal
Emotion associated with the intention to exercise was assessed
immediately following each measurement of intention. Partici-
pants were asked to rate their feelings when thinking about their
intention to exercise the following day. They rated four emotional
adjectives (that were selected from 24 adjectives of the Profile of
 C. Mohiyeddini et al. / Psychology of Sport and Exercise 10 (2009) 226–234
Mood States-Adolescents; Lane & Terry, 1998) on a 6 point scale
ranging from 1 (not at all) to 6 (extremely). Initially, three experts
rated the 24 adjectives with respect to their usefulness to assess
potential emotions associated with the intention to exercise. The
eight adjectives with the highest ratings (active, alert, energetic,
lively, downhearted, unhappy, uncertain, worried) were then pre-
sented to 30 undergraduate psychology students. They were asked
to imagine that they intended to exercise regularly. The eight
adjectives were then rated in terms of their usefulness in capturing
the emotional responses associated with the intention to exercise
and the four items with the highest ranking were selected (alert,
energetic, unhappy, and downhearted).
Physical activity: frequency and duration
Following every exercise session, participants completed an
Internet-based electronic exercise diary. Participants were asked
about the type of exercise, location and duration. Participants were
required to answer all the questions, as blank answers were not
permitted in any of the questions. Exercise frequency is considered
to be an indicator of maintenance of physical activity.
Procedure
The study was conducted over the course of nine weeks.
Following their response to a recruitment email, participants
received a postal questionnaire pack (including prepaid return
envelope). The participants were requested to mark the question-
naires with a code word to enable data integration with intention,
emotion and exercise behaviour measurement at a later point. The
questionnaires included requests for demographic and socioeco-
nomic information as well as the attitude, subjective norm and PBC
scales described above. One week later, participants were emailed
a username and password to enable them to access their personal
exercise diaries online. In accordance with the recommendations of
the U.S. Surgeon General, the American College of Sports Medicine
and the Centers for Disease Control and Prevention, participants
were asked to exercise twice a week for 30–60 min (USDHHS, 1996).
At the beginning of the first of eight weeks designated for exercising,
participants each nominated two days on which they were able to
exercise for 30 min. On the evening prior to their first nominated
day, they were sent a reminder email with a request to log on to their
personal exercise diary in order to complete intention and emotion
ratings with respect to their first planned exercise session. The
following day, participants were required to log on to fill out the
electronic post-exercise questionnaire, to indicate whether they
had completed their planned training and, if yes, how long their
exercise session had lasted. This process was repeated for the
remaining 15 exercise sessions (2 sessions per week over 8 weeks).
Participants were requested to continue completing the electronic
intention and emotion scales for subsequent exercise sessions even
if they had missed one or more sessions. Participants were offered
individual feedback as an incentive for taking part in this study.
Data reduction and analysis
In order to investigate the effects of attitudes, subjective norms, PBC
and intention on training duration and training frequency (model 1, see
Fig. 1) structural equation modelling (SEM) using LISREL 8.8 was per-
formed (Jöreskog & Sörbom, 1999). Using the same statistical proce-
dure, emotion was then included in the model 2 (Fig. 2) to examine the
effects of the predictor variables on duration and frequency of training
and also the mediating effect of emotion on duration and frequency of
training. Prior to analysis, all variables were normalised using the
PRELIS 2.50 algorithm (Jöreskog & Sörbom, 1999), as the assumption of
multivariate normality was violated (skewness and kurtosis in the
univariate analysis contradicted normality).
229
Model specification
In accordance with TPB, four latent variables were specified:
attitude (6 items), subjective norms (3 items) and behavioural
control (4 items). The intention variable was calculated as the mean
of the two items (mean correlation r ¼ 0.55; p < 0.01, SD ¼ 0.16)
relating to intention at each point of measurement. The average
correlation of the 16 intention measures was r ¼ 0.45 (p < 0.01,
SD ¼ 0.13). The intention measures were then pooled across two
week periods resulting in four indicators for measurement of the
latent variable ‘intention’ (intention indicators I1–I4). Cronbach’s
alpha for these four indicators was 0.91. The average frequency and
duration of physical activity were included as dependent variables.
Taking into account the recommendations of Dunlap and
Kemery (1987) for the control of multicollinearity, the latent vari-
able ‘emotion’ was specified as follows for the extended model.
Firstly, negative emotion items were recoded. At each time point,
the four emotion indicators correlated highly (mean correlation
r ¼ 0.64; range ¼ 0.52–0.81, SD ¼ 0.09). In a first step the four
emotion indicators were pooled for each assessment point result-
ing in 16 time specific measures. The average correlation of these
was r ¼ 0.62 (p < 0.01, SD ¼ 0.06). In a second step, analogous to the
calculation of the latent intention variable, emotion measures were
pooled across two week periods in order to obtain four indicators of
emotion (emotion indicators E1–E4). These four indicators yielded
a Cronbach’s alpha of 0.96.
Method of estimation and fit indices
SEM was performed using the method of maximum likelihood
(FML) estimation (see, e.g. Bollen, 1989). FML is the most widely used
fitting function for general structural equation models. Usually FML
is scale invariant and scale free (Swaminathan & Algina, 1978).
Lisrel provides a range of goodness-of-fit indices to estimate the
adequacy of the proposed model. The most common test for the
assessment of model fit is the chi square goodness-of-fit test (c2).
Since the c2 statistic is highly sensitive to sample size and often
inflates Type 1 error (Bollen, 1989), it is now recommended to
employ a combination of fit indices in conjunction with the c2
statistic to determine the adequacy of model fit (Bentler, 1990; Hu &
Bentler, 1999; Jöreskog, 1993; Jöreskog & Sörbom, 1993; Schu-
macker & Lomax, 1996): the standardized root-mean-square
residual (SRMR), the comparative fit index (CFI) and the root-mean
square error of approximation (RMSEA) are used as indicators of
model misspecification (Hu & Bentler, 1998). Hu and Bentler (1998)
suggest that cut-off values of SRMR  0.08, RMSEA  0.08 and
CFI  0.95 indicate a relatively good fit between the hypothesised
model and the observed data. Another minimum sample discrep-
ancy function, the c2/df ratio, is considered to be a useful criterion.
Bollen and Long (1993) suggest a c2/df ratio not larger than 2–5
times the degrees of freedom.
Results
Training duration and frequency
In total, participants completed 2670 exercise sessions
(M ¼ 11.27, SD ¼ 4.23) which equals 70% of the foreseen sessions. The
average training duration was 32.11 min (SD ¼ 7.83), thus indicating
that, on average, participants achieved the minimum prescribed
exercise duration of 30–60 min per session. Table 1 shows the
correlations, means and the standard deviation of all variables.
Model 1: the prediction of training frequency and duration:
traditional TPB model
Table 2 shows the factor loadings and internal consistency
coefficients (Cronbach’s alpha) for all latent variable indicators.
230 C. Mohiyeddini et al. / Psychology of Sport and Exercise 10 (2009) 226–234

Attitude

.18* (.19*)
.16

Frequency
Subjective .20** (.19*) Intention .70** (.65**)
.22** (duration)
Norm 2
R = .17 (.17)
R2= .49 (.42)

.18*
.22** (.22**) .02 (.03)

Perceived
Behavioural
Control

Note. *p < 0.05; ** p < 0.01.

Fig. 1. Model 1 with standardized regression coefficients.

For frequency of training, the traditional hypothesised predic-
tion model (Fig. 1) yielded the following fit indices: c2(128,
N ¼ 237) ¼ 196.60, p ¼ 0.000, RMSEA ¼ 0.048 (90% CI ¼ 0.034;
0.060), p(RMSEA < 0.05) ¼ 0.60, CFI ¼ 0.97, NNFI ¼ 0.96,
SRMR ¼ 0.063 and c2/df ¼ 1.53. RMSEA, CFI and NNFI are above the
acceptable cut-off (0.95) and the c2/df ratio is very small. Hence the
model can be accepted. Intention has a substantial impact on
frequency of training (B ¼ 0.70, t(235) ¼ 12.21, p < 0.001) and can
explain 49% of variance in exercise frequency: stronger intention
predicts higher frequency of the training session indicating better
maintenance. Attitudes (B ¼ 0.18, t(235) ¼ 2.36, p < 0.05), subjec-
tive norms (B ¼ 0.20, t(235) ¼ 2.80, p < 0.01) and PBC (B ¼ 0.22,
t(235) ¼ 2.94, p < 0.01) all have a significant impact on intention
and explain 17% of the variance in intention. There was no direct
effect of PBC on frequency of training (B ¼ 0.02, t(235) ¼ 0.42, n.s.).
Therefore the link between PBC and frequency of training was
suppressed in model 2.
The initial traditional prediction model (Fig. 1) showed the
following fit indices for the duration of training: c2(128,
N ¼ 237) ¼ 186.57, p ¼ 0.000, RMSEA ¼ 0.044 (90% CI ¼ 0.029;
0.057), p(RMSEA < 0.05) ¼ 0.76, CFI ¼ 0.97, NNFI ¼ 0.97,
SRMR ¼ 0.063 and c2/df ¼ 1.45. CFI and NNFI are clearly above the
acceptable cut-off (0.95). Furthermore the very small size of the c2/
df ratio indicates also that the model has a good fit. RMSEA, and
SRMR are small which suggested that discrepancies between the
present model and the true model were relatively small. Globally,

Emotion
R2= .27 (.27)
Attitude

.16 .19*(.19*)
.52** (.52**) .45** (.53**)

Frequency
Subjective Intention
.22** .48** (.37**) (duration)
Norm .20** (.19*)
R2= .17 (.17) R2= .66 (.62)

.18*
.21** (22**)
Perceived
Behavioural
Control

Note. * p < 0.05; ** p < 0.01.

Fig. 2. Model 2 with standardized regression coefficients.

 C. Mohiyeddini et al. / Psychology of Sport and Exercise 10 (2009) 226–234 231

Table 1
Correlations, means and standard deviations of all variables

A1 A2 A3 A4 A5 A6 SN1 SN2 SN3 PBC1 PBC2 PBC3 PBC4 I1 I2 I3 I4 E1 E2 E3 E4 D F

A1 –
A2 0.21** –
A3 0.38** 0.26** –
A4 0.38** 0.28** 0.28** –
A5 0.29** 0.30** 0.42** 0.18** –
A6 0.34** 0.40** 0.43** 0.34** 0.29** –
SN1 0.19** 0.01 0.10 0.12 0.01 0.13* –
SN2 0.11 0.06 0.04 0.10 0.00 0.05 0.75** –
SN3 0.15* 0.01 0.10 0.21** 0.09 0.13 0.57** 0.58** –
PBC1 0.26** 0.03 0.15* 0.12 0.06 0.05 0.14* 0.14* 0.17** –
PBC2 0.26** 0.04 0.13* 0.09 0.06 0.05 0.14* 0.11 0.14* 0.58** –
PBC3 0.22** 0.04 0.16* 0.14* 0.05 0.06 0.09 0.05 0.08 0.59** 0.57** –
PBC4 0.21** 0.01 0.03 0.14* 0.03 0.10 0.12 0.09 0.06 0.50** 0.55** 0.53** –
I1 0.20** 0.04 0.18** 0.03 0.01 0.16* 0.19** 0.16* 0.16* 0.18** 0.20** 0.26** 0.09 –
I2 0.27** 0.16* 0.24** 0.03 0.05 0.17** 0.22** 0.18** 0.16* 0.20** 0.23** 0.26** 0.09 0.80** –
I3 0.10 0.08 0.16* 0.03 0.04 0.10 0.12 0.07 0.04 0.06 0.13* 0.17* 0.02 0.69** 0.73** –
I4 0.23** 0.08 0.19** 0.03 0.00 0.19** 0.30** 0.28** 0.23** 0.23** 0.27** 0.33** 0.09 0.74** 0.73** 0.68** –
E1 0.19** 0.03 0.14* 0.05 0.09 0.02 0.13* 0.08 0.08 0.14* 0.16* 0.13 0.12 0.39** 0.41** 0.32** 0.32** –
E2 0.18** 0.05 0.11 0.07 0.09 0.02 0.15* 0.11 0.08 0.05 0.10 0.07 0.03 0.35** 0.37** 0.28** 0.30** 0.89** –
E3 0.24** 0.10 0.20** 0.10 0.06 0.05 0.12 0.07 0.08 0.17* 0.19** 0.20** 0.12 0.52** 0.52** 0.42** 0.45** 0.87** 0.86** –
E4 0.27** 0.07 0.16* 0.11 0.08 0.03 0.19** 0.13* 0.15* 0.13* 0.16* 0.14* 0.06 0.47** 0.50** 0.39** 0.43** 0.86** 0.88** 0.88** –
D 0.19** 0.08 0.12 0.06 0.09 0.06 0.19** 0.13* 0.12 0.10 0.14* 0.16* 0.07 0.58** 0.59** 0.52** 0.50** 0.67** 0.65** 0.71** 0.68** –
F 0.20** 0.08 0.12 0.00 0.09 0.07 0.26** 0.15* 0.12 0.15* 0.20** 0.22** 0.11 0.61** 0.63** 0.61** 0.59** 0.64** 0.64** 0.68** 0.66** 0.78** –
M 4.59 4.43 4.43 4.41 4.37 4.47 4.58 4.70 4.19 4.37 4.31 4.31 4.32 3.87 4.05 2.86 2.79 3.44 3.25 3.26 3.36 32.11 11.27
SD 1.37 1.31 1.31 1.32 1.39 1.36 1.17 1.16 1.22 1.37 1.41 1.44 1.42 0.76 0.91 0.81 0.70 1.03 1.10 1.02 0.97 7.83 4.23

Note. N ¼ 237. *p < 0.05, **p < 0.01. A1–A6 ¼ attitude 1 to attitude 6, SN1–SN3 ¼ subjective norm 1 to subjective norm 3, PBC1–PBC4 ¼ perceived behavioural control 1 to
perceived behavioural control 4, I1–I4 ¼ Intention indicators 1–4, E1–E4 ¼ Emotion indicators 1–4. D ¼ Average duration of exercise. F ¼ Average frequency of exercise.

these fit indices suggest that the model is acceptable. Attitudes
(B ¼ 0.19, t(235) ¼ 2.41, p < 0.05), subjective norms (B ¼ 0.19,
t(235) ¼ 2.73, p < 0.01) and PBC (B ¼ 0.22, t(235) ¼ 2.55, p < 0.01)
all have a significant impact on intention and explain 17% of the
variance in intention. More positive attitudes and subjective norms
and more PBC predict stronger intentions. Intention has a strongly
significant impact on exercise duration (B ¼ 0.65, t(235) ¼ 11.23,
p < 0.001) and explains 42% of the variance of exercise duration –
participants with stronger intentions exercise longer. PBC has no
significant direct impact on exercise duration (B ¼ 0.03, t
(235) ¼ 0.56, n.s.). Therefore the link between PBC and exercise
duration was excluded in model 2.
Model 2: are the predictor effects on duration and frequency of
training mediated by emotions?
According to Baron and Kenny (1986) four conditions must be
met when a variable functions as a mediator variable: (1) the link
between the independent variable (intention) and the dependent
variable (frequency and duration of training) must be significant,
(2) the link between the independent variable and the mediator
variable (emotion) must be significant, (3) the link between the
mediator variable and the dependent variable must be significant,
and (4) testing if the mediator variable reduces or eliminates the
link between the independent variable and the dependent variable.

Table 2
Factor loadings for constructs

Frequency of training Duration of training

Factor loadings within Factor loadings within Factor loadings within Factor loadings within
constructs for model 1 constructs for model 2 constructs for model 1 constructs for model 2
A1 0.59 0.59 0.59 0.60
A2 0.52 0.52 0.52 0.51
A3 0.65 0.65 0.65 0.65
A4 0.52 0.52 0.52 0.52
A5 0.50 0.49 0.49 0.50
A6 0.65 0.65 0.65 0.65
SN1 0.87 0.87 0.87 0.87
SN2 0.86 0.86 0.86 0.86
SN3 0.66 0.66 0.67 0.67
PBC1 0.76 0.76 0.76 0.76
PBC2 0.77 0.77 0.77 0.77
PBC3 0.77 0.77 0.77 0.77
PBC4 0.68 0.68 0.68 0.68
I1 0.88 0.88 0.88 0.88
I2 0.90 0.90 0.90 0.90
I3 0.80 0.80 0.80 0.80
I4 0.83 0.83 0.83 0.83
E1 – 0.93 – 0.93
E2 – 0.94 – 0.94
E3 – 0.93 – 0.93
E4 – 0.94 – 0.94

Note. A1–A6 ¼ attitude 1 to attitude 6, SN1–SN3 ¼ subjective norm 1 to subjective norm 3, PBC1–PBC4 ¼ perceived behavioural control 1 to perceived behavioural control 4,
I1–I4 ¼ Intention indicators 1–4, E1–E4 ¼ Emotion indicators 1–4.
232 C. Mohiyeddini et al. / Psychology of Sport and Exercise 10 (2009) 226–234
While condition 1 was tested in model 1, conditions 2, 3, and 4 are
investigated in model 2.
The hypothesised mediation model 2 (Fig. 2) shows the
following fit indices for the training frequency: c2(201,
N ¼ 237) ¼ 343.85, p ¼ 0.00, RMSEA ¼ 0.055 (90% CI ¼ 0.044;
0.065), p(RMSEA < 0.05) ¼ 0.20, CFI ¼ 0.97, NNFI ¼ 0.97,
SRMR ¼ 0.061 and c2/df ¼ 1.71. The fit indices indicate that the
mediation model can be accepted. Intention (B ¼ 0.48,
t(235) ¼ 9.26, p < 0.001) and emotions (B ¼ 0.45, t(235) ¼ 9.14,
p < 0.001) have significant impact on the frequency of training and
can explain 66% of variance in exercise frequency – stronger
intention and more positive emotions lead to higher frequency of
the training sessions. In addition, intention has a substantial impact
on (positive) emotions (B ¼ 0.52, t(235) ¼ 8.09, p < 0.001), stronger
intention predicts a higher level of positive emotions.
With respect to duration of training the hypothesised mediation
model 2 (Fig. 2) generated the following fit indices: c2(201,
N ¼ 237) ¼ 345.90, p ¼ 0.00, RMSEA ¼ 0.055 (90% CI ¼ 0.045;
0.065), p(RMSEA < 0.05) ¼ 0.19, CFI ¼ 0.97, NNFI ¼ 0.97,
SRMR ¼ 0.064 and c2/df ¼ 1.72. According to these fit indices the
mediation model can be accepted. Intention (B ¼ 0.37, t(235) ¼ 7.06,
p < 0.001) and emotions (B ¼ 0.53, t(235) ¼ 10.26, p < 0.001) have
a substantial impact on the duration of training and can explain 62%
of variance in exercise duration – stronger intention and more
positive emotions predict longer duration of training sessions.
Furthermore intention has a significant impact on (positive)
emotions (B ¼ 0.52, t(235) ¼ 8.16, p < 0.001) – stronger intention
predicts a higher level of positive emotions.
Further exploratory analysis
In two further analyses (for duration and frequency) we tested
the question whether an extension of model 2 in which regression
paths of attitudes, subjective norms, and PBC on emotions are free,
lead to a better model fit and an increase of the explained variance
of emotions. The results show that none of the regression paths are
significant (Bs < 0.05, ts(235) < 0.74, n.s.).
Discussion
The aim of this research was to examine predictors of duration
and maintenance of physical activity in the context of the TPB. The
central assumption was that inclusion of an emotion variable as
a mediator between intention and behaviour may serve to diminish
the ‘intention–behaviour gap’ which has been commonly observed
in studies based on TPB (Hagger et al., 2002).
The initial model tested in this research was a traditional TPB
based model. The results demonstrated that the intention compo-
nent of the TPB could account for the duration and maintenance of
exercise behaviour. Furthermore findings indicate that satisfactory
prediction of intention to exercise can be achieved from attitudes,
subjective norms and perceived behavioural control (PBC). Whilst
the predictor variables account only for 17% of the variance in
intention to exercise, a low to moderate amount of explained
variance in comparison to the studies cited in Hagger et al.’s (2002)
meta-analysis, all three predictors contributed significantly to the
prediction of intention to exercise as expected. Furthermore, 49% of
the variance in training frequency and 42% in training duration
were accounted for by the measure of intention to exercise. These
represent relatively high proportions of variance in accounting for
the relationship between intention and behaviour compared to
Sutton (1998). One reason for this may be that the design of the
study required participants to make specific plans by rating their
intention to exercise and emotions associated with the intention
prior to each training episode. This supports Sniehotta, Scholz, and
Schwarzer’s (2005) finding that variables such as planning may
lead to an increase of the explained variance of the TPB model.
Moreover, in contrast with Sheeran, Conner, and Norman (2001)
but in line with Armitage (2005), the present study supports the
idea that the TPB can account for the maintenance of health
behaviour. Consistent with prior retrospective studies, the present
research demonstrated that intention was the dominant predictor
of exercise maintenance (measured as frequency of exercise
behaviour). This finding differs from the results of Armitage (2005),
which indicated that PBC was the dominant predictor of exercise
maintenance. A possible explanation for this is that Armitage’s
sample reported a high initial level of PBC (M ¼ 6.02, SD ¼ 1.05) and
relatively low intention scores (M ¼ 2.49, SD ¼ 0.80). In contrast,
the sample in the present study displayed substantially lower PBC
scores (M ¼ 4.32, SD ¼ 1.14) and higher intention (M ¼ 3.04,
SD ¼ 0.96). Another reason for this discrepancy may lie in the fact
that attitudes, subjective norms and PBC were measured only at
baseline in the present study to minimise the burden on partici-
pants. This lack of temporal correspondence may provide an invalid
test of the mediation effect of emotions. Furthermore, the attitude
items did not correspond with the intention items (twice a week of
exercise was specified only in the attitude items) which resulted in
less variance being explained in intentions as compared to behav-
iour. Concurrently, the close temporal proximity of intention and
emotion ratings and the observed exercise behaviour might have
inflated the correspondence between intention, emotion and
behaviour. This limitation of the current study should be addressed
in future research. However, it seems unlikely that the significant
predictive validity of emotions (adding around 17% to the predic-
tion of behaviour) would become non-significant if PBC and
behaviour corresponded in the element of time.
The hypothesis that emotions associated with the intention to
exercise mediate the intention–behaviour relationship was sup-
ported both with respect to exercise frequency and duration. Baron
and Kenny’s conditions for a mediation model are, thus, fulfilled:
(1) intention predicts frequency and duration of the training
sessions, (2) intention and emotions are highly related, (3) emotion
has a substantial impact on both exercise duration and frequency,
and (4) the mediator variable ‘emotion’ reduces the link between
intention and the frequency and duration of training significantly
(partial mediation). Inclusion of the emotion variable resulted in
a 17% increase in variance explained in exercise frequency and
a 20% increase in predicting exercise duration. The observed
increase in variance accounted for in behaviour is accompanied by
a reduction in strength of the direct link between intention and
behaviour (22% decrease for exercise frequency and 28% for exer-
cise duration). This partial mediation effect of emotion emphasises
on the one hand the role of intention as a main predictor of
behaviour and clarifies on the other hand that in a ‘‘theory deep-
ening’’ procedure (Perugini & Bagozzi, 2001, p. 80) emotion as
a mediator variable can improve the predictive validity of the
traditional TPB model.
These findings fully replicate those of Mohiyeddini and Bauer
(2007), but the observed effects are considerably stronger in the
current study. Thus, the longitudinal, prospective design incorpo-
rating repeated measurement of both intentions and emotions
associated with intentions has resulted in explanation of a large
proportion of the variance in the intention–behaviour relationship.
The increase in explained variance in the mediator model raises
the question of the theoretical mechanisms by which emotion is
influenced by intention on the one hand and affects behaviour on
the other. Differences may arise, for example, when intentions are
motivated by approach (e.g. positive aspect of physical activity) or
avoidance (e.g. the negative consequence of not being active)
tendencies respectively (Carver et al., 2000). In the present study
emotion was measured purely as positive or negative affect asso-
ciated with the intention to exercise. A more differentiated
 C. Mohiyeddini et al. / Psychology of Sport and Exercise 10 (2009) 226–234
assessment of emotional appraisal of reported intentions may yield
a differential account of the motivational factors underlying such
reported emotions. This may be further influenced by self-regula-
tion and emotion regulation variables (Mohiyeddini, 2005). Thus it
might be expected that hedonistic emotion regulation (the ability
to maintain and intensify positive emotions and to terminate
negative emotions) strengthens positive emotions and weakens
negative emotions associated with intention. On the other hand
distress-augmenting emotion regulation (maintenance and inten-
sification of negative emotion) should have the opposite effect.
With respect to the link between emotion and behaviour, it is
not clear how or when emotions affect a behavioural sequence.
Initiation, maintenance, completion or re-initiation of behaviour
(Lewis & Haviland, 1993) may be differentially implicated. Similarly,
Zhu and Thagard (2002) have hypothesised that emotions may
impact action generation, action execution and/or action control.
However, further research needs to clarify the behavioural and
emotional processes underlying the predictive power of positive
emotions. Furthermore, future research should consider that
emotions influence availability of mental capacity producing
‘‘ironic effects’’ (Wegner, 1994). As Wegner pointed out under
conditions of stress (e.g. produced by positive or negative
emotions) people tend to generate the opposite of what they intend
to generate so that ‘‘desired happiness becomes sadness, desired
relaxation becomes anxiety, desired interest becomes boredom,
desired love becomes hate, and so on’’ (p. 39; see also Wegner,
Erber, & Zanakos, 1993). In addition, possible moderator effects, i.e.
an interaction effect of intention and emotion on exercise behav-
iour should be investigated.
Given that we used a non-random sampling method and
participation required IT availability, the external validity of the
present study is under question. However, since both the SES
variables and previous experience with training did not show any
effects, the results of the study may apply to the general population
of Switzerland.
Several limitations of this study should be addressed in future
research. First, exclusively self-report measures were used. Self-
reports require the ability to accurately reflect on and report one’s
behaviour. They also rely on recall, are susceptible to memory and
may reflect a unique pattern of reporting biases (e.g. primacy and
recency effects; Murdock, 1962) and social desirability (e.g. report
of incorrect duration or frequency of training). Also, the Likert type
items employed in this study for the purpose of measuring the
valence and intensity of emotions associated with the intention to
exercise may be regarded as a cognitively mediated emotion
measurement (see Robinson & Clore, 2002). Thus ratings of such
items represent cognitive appraisal of the underlying emotion. A
more direct assessment of emotion might be obtained by using, for
example physiological measures of arousal during contemplation of
the intention. However, whilst physiological measures may be
more informative in terms of the intensity of arousal, they are less
clear with respect to the valence of the associated emotions.
Furthermore, in model 2 positive and negative emotions were
represented as a single latent variable to avoid multicollinearity
between a latent variable ‘‘positive emotions’’ and a latent variable
‘‘negative emotions’’. Although this ensures parsimony, the main
issue with this strategy is, as indicated by previous research, that
positive emotions do not necessarily reflect a lack of negative
emotions (Watson & Clark, 1984). However, further research is
needed to explore the differential effects of positive and negative
emotions as mediators of the intention–behaviour link. A further
limitation is a lack of consideration of the intensity of completed
exercise sessions. Including measures of training intensity would
facilitate a more detailed interpretation of frequency and duration
of exercise behaviour. Given the potential pitfalls of self-reporting
and the crucial role of exercise behaviour, it is essential that we
233
thoroughly examine the role of emotions in bridging the intention–
behaviour gap by using objective measures of exercise behaviour in
future research.
In sum, the results emphasise the relevance of TPB in the context
of exercise behaviour. In addition, our study points to an important
theoretical possibility of how to increase the predictive power of
the TPB. As Armitage (2005) pointed out, research into mainte-
nance of exercise behaviour as a process is still in its infancy. One
potential application of our results for research into maintenance
would be to describe not only the content of an intervention and its
duration but also the emotional processes and their changes
occurring during the maintenance of health behaviour in general,
and exercise behaviour specifically. We suggest considering both
emotional processes as well as emotion regulation processes
(Gross, 2007) in future research and in the design and imple-
mentation of intervention programs. Specifically an intervention
focusing on hedonistic emotion regulation (Mohiyeddini, 2005;
Wirtz et al., 2006) may facilitate the implementation of desired
health behaviours and contribute to their maintenance over time.
References
Abraham, C., Sheeran, P., & Johnston, M. (1998). From health beliefs to self-regu-
lation: theoretical advances in the psychology of action control. Psychology and
Health, 13, 569–591.
Ajzen, I. (1991). The theory of planned behavior. Organisational Behavior and Human
Decision Processes, 50, 179–211.
Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social
behavior. Englewood Cliffs, NJ: Prentice Hall.
Armitage, C. J., & Conner, M. (2000). Social cognition models and health behavior:
a structured review. Psychology and Health, 15, 173–189.
Armitage, J. (2005). Can the theory of planned behavior predict the maintenance of
physical activity? Health Psychology, 3, 235–245.
Armstrong, N., & McManus, A. (1994). Children’s fitness and physical activity:
a challenge for physical education. British Journal of Physical Education, 25,
20–26.
Bagozzi, P. B., Moore, D. J., & Leone, L. (2004). Self-control and the self-regulation of
dieting decisions: the role of prefactual attitudes, subjective norms, and resis-
tance to temptation. Basic and Applied Social Psychology, 26, 199–213.
Bagozzi, R. P. (1981). Attitudes, intentions and behavior: a test of some key
hypotheses. Journal of Personality and Social Psychology, 41, 607–627.
Bagozzi, R. P. (1992). The self-regulation of attitudes, intentions, and behavior. Social
Psychology Quarterly, 55, 178–204.
Bagozzi, R. P., Baumgartner, H., & Pieters, R. (1998). Goal-directed emotions.
Cognition and Emotion, 12, 1–16.
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in
social psychological research: conceptual, strategic, and statistical consider-
ations. Journal of Personality and Social Psychology, 51, 1173–1182.
Bentler, P. M. (1990). Comparative fit indices in structural models. Psychological
Bulletin, 107, 238–246.
Blanchard, C., Rodgers, W., Courneya, K., Daub, B., & Knapik, G. (2002). Does barrier
efficacy mediate the gender/exercise adherence relationship during phase II
cardiac rehabilitation? Rehabilitation Psychology, 47, 106–120.
Bollen, K. A. (1989). Structural equations with latent variables. New York: Wiley.
Bollen, K., & Long, J. (Eds.). (1993). Testing structural equation models. Thousand
Oaks, CA: Sage.
Brown, A., & Siahpush, M. (2006). Socioeconomic predictors of a sedentary lifestyle:
results from the 2001 National Health Survey. Journal of Physical Activity and
Health, 3, 90–101.
Carver, C., Sutton, S. K., & Scheier, M. F. (2000). Action, emotion, and personality:
emerging conceptual integration. Personality and Social Psychology Bulletin, 26,
741–751.
Chatzisarantis, N., & Hagger, M. (2005). Effects of a brief intervention based on the
theory of planned behaviour on leisure time physical activity participation.
Journal of Sport and Exercise Psychology, 27, 470–487.
Chen, M., & Bargh, J. A. (1999). Consequences of automatic evaluation: immediate
behavioral predispositions to approach and avoid the stimulus. Personality and
Social Psychology Bulletin, 25, 215–224.
Conner, M., & Armitage, C. J. (1998). Extending the theory of planned behavior:
a review and avenues for further research. Journal of Applied Social Psychology,
28, 1429–1464.
Coumeya, K. S., & McAuley, E. (1994). Are there different determinants of the
frequency, intensity, and duration of physical activity? Behavioral Medicine, 20,
84–90.
Dillon, W. R., & Kumar, A. (1985). Attitude organization and the attitude–behavior
relation: a critique of Bagozzi and Burnkrant’s reanalysis of Fishbein and Ajzen.
Journal of Personality and Social Psychology, 49, 33–46.
Dishman, R. K., & Buckworth, J. (2001). Exercise psychology. Champaign, IL: Human
Kinetics.
234 C. Mohiyeddini et al. / Psychology of Sport and Exercise 10 (2009) 226–234
Dunlap, W. P., & Kemery, E. R. (1987). Failure to detect moderating effects: is
multicollinearity the problem? Psychology Bulletin, 102, 418–420.
Eagly, A. H., & Chaiken, S. (1993). The psychology of attitudes. Belmont, CA:
Wadsworth.
Estabrooks, P., & Carron, A. V. (1998). The conceptualization and effect of control
beliefs on exercise attendance in the elderly. Journal of Aging and Health, 10,
441–457.
Gross, J. J. (Ed.). (2007). Handbook of emotion regulation. New York: Guilford Press.
Hagger, M. S., Chatzisarantis, N., Biddle, S., & Orbell, S. (2001). Antecedents of
children’s physical activity intentions and behavior: predictive validity and
longitudinal effects. Psychology and Health, 16, 391–407.
Hagger, M. S., Chatzisarantis, N. L. D., & Biddle, S. J. H. (2002). A meta-analytic
review of the theories of reasoned action and planned behavior in physical
activity: predictive validity and the contribution of additional variables. Journal
of Sport, & Exercise Psychology, 24, 3–32.
Henderson, K. A., Bialeschki, M. D., Shaw, S. M., & Freysinger, V. J. (1999). Both gaps
and gains: Feminist perspectives on women’s leisure. State College, PA: Venture
Publishing.
Hu, L., & Bentler, M. (1998). Fit indices in covariance structure modeling: sensitivity
to underparameterized model misspecification. Psychological Methods, 3,
424–453.
Hu, L., & Bentler, M. (1999). Cut-off criteria for fit indexes in covariance structure
analysis: conventional criteria versus new alternatives. Structural Equation
Modeling: A Multidisciplinary Journal, 6, 1–55.
Jöreskog, K. (1993). Testing structural equation models. In K. A. Bollen, & J. S. Long
(Eds.), Testing structural equation models (pp. 307–314). Newbury Park: Sage
Publications.
Jöreskog, K. G., & Sörbom, D. (1993). Structural equation modeling with the SIMPLIS
command language. Chicago, IL: Scientific Software.
Jöreskog, K. G., & Sörbom, D. (1999). LISREL 8.51. Chicago: Scientific Software.
Lane, A. M., & Terry, P. C. (1998). Development of normative data for profile of mood
states-C among adult and young athletes. Journal of Sports Sciences, 16, 94–95.
Leventhal, H., & Scherer, K. (1987). The relationship of emotion to cognition:
a functional approach to a semantic controversy. Cognition and Emotion, 1, 3–28.
Lewis, M., & Haviland, J. M. (1993). Handbook of emotions. New York: Guilford Press.
Martinez-Gonzalez, M. A., Varo, J. J., Santos, J. L., De Irala, J., Gibney, M., Kearney, J., et al.
(2001). Prevalence of physical activity during leisure time in the European Union.
Medicine and Science in Sports and Exercise, 33, 1142–1146.
Marttila, J., & Nupponen, R. (2000). Health enhancing physical activity as perceived
in interviews based on the theory of planned behaviour. Psychology and Health,
15, 593–608.
Milne, S., Orbell, S., & Sheeran, P. (2002). Combining motivational and volitional
interventions to promote exercise participation: protection motivation theory
and implementation intentions. British Journal of Health Psychology, 7, 163–184.
Mohiyeddini, C. (2005). Persönlichkeit, Emotionsregulation und Psychopathologie.
[Personality, emotion regulation and psychopathology]. Habilitationsschrift.
Universität Tübingen.
Mohiyeddini, C., & Bauer, S. (2007). Intentions-Verhaltens-Lücke bei sportlichen
Aktivitäten: Die Bedeutung von Emotionen. [Intention–behavior gap in physical
exercise: the role of emotions]. Zeitschrift für Sportpsychologie, 14, 3–13.
Murdock, B. (1962). The serial position effect of free recall. Journal of Experimental
Psychology, 64, 482–488.
Pate, R. R., Dowda, M., O’Neill, J. R., & Ward, D. S. (2007). Change in physical activity
participation among adolescent girls from 8th to 12th grade. Journal of Physical
Activity, & Health, 4, 3–16.
Perugini, M., & Bagozzi, R. P. (2001). The role of desires and anticipated emotions in
goal-directed behavior: broadening and deepening the theory of planned
behavior. British Journal of Social Psychology, 40, 79–98.
Robinson, M. D., & Clore, G. L. (2002). Belief and feeling: evidence for an accessi-
bility model of emotional self-report. Psychological Bulletin, 128, 934–960.
Scherer, K. R., Schorr, A., & Johnstone, T. (Eds.). (2001). Appraisal processes in
emotion: Theory, method, research. New York: Oxford University Press.
Schumacker, R. E., & Lomax, R. G. (1996). A beginner’s guide to structural equation
modeling. Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.
Sheeran, P. (2002). Intention–behavior-relations: a conceptual and empirical
review. In M. Hewstone, & W. Stroebe (Eds.), European review of social
psychology, Vol. 12 (pp. 1–36). New York: Wiley.
Sheeran, P., Conner, M., & Norman, P. (2001). Can the theory of planned behavior
explain patterns of health behavior change? Health Psychology, 20, 12–19.
Sheeran, P., & Orbell, S. (1998). Do intentions predict condom use? Meta-analysis
and examination of six moderator variables. British Journal of Social Psychology,
37, 231–250.
Sniehotta, F. F., Scholz, U., & Schwarzer, R. (2005). Bridging the intention–behavior
gap: planning, self-efficacy, and action control in the adoption and maintenance
of physical exercise. Psychology, & Health, 20, 143–160.
Solomon, R. C. (2000). The philosophy of emotions. In M. Lewis, & J. M. Haviland-
Jones (Eds.), Handbook of emotions (2nd ed.). (pp. 17–29) New York: Guilford.
Sutton, S. (1998). Predicting and explaining intentions and behavior: how well are
we doing? Journal of Applied Social Psychology, 28, 1317–1338.
Swaminathan, H., & Algina, J. (1978). Scale freeness in factor analysis. Psychometrika,
43, 581–583.
U.S. Department of Health and Human Services. (1996). Physical activity and health:
A report to the surgeon general. Atlanta, GA: U.S. Department of Health and
Human Services, Centers for Disease Control and Prevention, National Center
for Chronic Disease Prevention and Health Promotion.
Van der Pligt, J., Zeelenberg, M., van Dijk, W. W., de Vries, N. K., & Richard, R. (1998).
Affect, attitudes and decisions: let’s be more specific. In W. Stroebe, &
M. Hewstone (Eds.), European review of social psychology (pp. 33–66). Chi-
chester: Wiley.
Wallston, K., & Armstrong, C. (2002). Theoretically-based strategies for health
behavior change. In M. P. O’Donnell (Ed.), Health promotion in the workplace (3rd
ed.). (pp. 182–201) Albany, NY: Delmar.
Watson, D., & Clark, L. A. (1984). Negative affectivity: the disposition to experience
aversive emotional states. Psychological Bulletin, 96, 465–490.
Wegner, D. M. (1994). Ironic processes of mental control. Psychological Review, 101,
34–52.
Wegner, D. M., Erber, R., & Zanakos, S. (1993). Ironic processes in the mental control
of mood and mood-related thought. Journal of Personality and Social Psychology,
65, 1093–1104.
Wirtz, P. H., Von Känel, R., Mohiyeddini, C., Emini, L., Ruedisueli, K., Groessbauer, S.,
et al. (2006). Low social support and poor emotional regulation are associated
with increased stress hormone reactivity to mental stress in systemic hyper-
tension. Journal of Clinical Endocrinology and Metabolism, 91, 3857–3865.
Zhu, J., & Thagard, P. (2002). Emotion and action. Philosophical Psychology, 15, 19–36.