Health Education & Behavior http://heb.sagepub.

com/

Positive Exercise Experience Facilitates Behavior Change via Self-Efficacy

Linda Parschau, Lena Fleig, Lisa Marie Warner, Sarah Pomp, Milena Barz, Nina Knoll, Ralf Schwarzer and Sonia Lippke
Health Educ Behav published online 10 April 2014
DOI: 10.1177/1090198114529132

The online version of this article can be found at:

http://heb.sagepub.com/content/early/2014/04/09/1090198114529132

Published by:

http://www.sagepublications.com

On behalf of:

Society for Public Health Education

Additional services and information for Health Education & Behavior can be found at:

Email Alerts: http://heb.sagepub.com/cgi/alerts

Subscriptions: http://heb.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

Downloaded from heb.sagepub.com at University of Waikato Library on May 25, 2014

529132
research-article2014
HEBXXX10.1177/1090198114529132Health Education & BehaviorParschau et al.

Article
Health Education & Behavior

Positive Exercise Experience 1­–9

© 2014 Society for Public
Health Education
Facilitates Behavior Change via Reprints and permissions:
sagepub.com/journalsPermissions.nav

Self-Efficacy DOI: 10.1177/1090198114529132

heb.sagepub.com

Linda Parschau, PhD1, Lena Fleig, PhD1, Lisa Marie Warner, PhD1,2,
Sarah Pomp, PhD1, Milena Barz, Dipl-Psych1, Nina Knoll, PhD1,
Ralf Schwarzer, PhD3,1, and Sonia Lippke, PhD4

Abstract
Purpose. Motivational processes can be set in motion when positive consequences of physical exercise are experienced.
However, relationships between positive exercise experience and determinants of the motivational and the volitional phases
of exercise change have attracted only sparse attention in research. Method. This research examines direct and indirect
associations between positive experience and motivational as well as volitional self-efficacy, intention, action planning, and
exercise in two distinct longitudinal samples. The first one originates from an online observational study in the general
population with three measurement points in time (N = 350) and the second one from a clinical intervention study in a
rehabilitation context with four measurement points (N = 275). Results. Structural equation modeling revealed the following:
Positive experience is directly related with motivational self-efficacy as well as intentions in both samples. In the online sample
only, positive experience is associated with volitional self-efficacy. In each sample, experience is indirectly associated with
action planning via motivational self-efficacy and intentions. Moreover, action planning, in turn, predicts changes in physical
exercise levels. Conclusions. Findings suggest a more prominent role of positive experience in the motivational than in the
volitional phase of physical exercise change. Thus, this research contributes to the understanding of how positive experience
is involved in the behavior change process.

Keywords
exercise, intention, motivation, phase-specific self-efficacy, volition

Despite the physical and mental health benefits of physical
activity (Saxena, Van Ommeren, Tang, & Armstrong, 2005;
Warburton, Nicol, & Bredin, 2006), many adults remain
physically inactive. According to the World Health
Organization (2011), 44.5% of the population in the indus-
trialized countries does not meet the recommendations of
being physically active for at least 30 minutes on 5 or more
days a week.
Lack of time, motivation, and energy as well as negative
experience with physical activity (e.g., sweating, pain)
seem to be main reasons for inactivity (Toscos, Consolvo,
& McDonald, 2011). However, a recent longitudinal twin
study (Aaltonen et al., 2012) concludes that perceived
­barriers are not as important for the initiation and mainte-
nance of regular physical activity as expected because no
difference in perceived barriers between active and inactive
co-twins was identified. Instead, experiencing positive con-
sequences such as pride as well as beneficial health effects
were found to be more conducive for regular activity.
Furthermore, the experience of well-being (feeling better
afterward), social aspects (e.g., meeting people during
exercise) and positive impact on one’s own appearance sup-
port engagement in physical activity (Sherwood & Jeffery,
2000; Taylor et al., 1999).
So far, the differential role of positive experience with exer-
cise for motivational and volitional processes has attracted
only sparse attention in research. However, a deeper under-
standing of how positive experience is related to physical
1
Freie Universität Berlin, Berlin, Germany
2
German Centre of Gerontology, Berlin, Germany
3
Institute for Positive Psychology and Education, Australian Catholic
University, Sydney, Australia
4
Jacobs University Bremen, Bremen, Germany
Corresponding Author:
Linda Parschau, Department of Health Psychology, Freie Universität
Berlin, Habelschwerdter Allee 45, 14195 Berlin, Germany.
Email: linda.parschau@fu-berlin.de

Downloaded from heb.sagepub.com at University of Waikato Library on May 25, 2014

2 Health Education & Behavior 
exercise as well as to intentions to perform health behaviors is
required to develop effective interventions.
In the course of health behavior change, different cogni-
tive processes have been found important. Most theories dis-
tinguish at least two phases—a motivational and a volitional
phase (Heckhausen, 1991). The motivational phase refers to
the formation of a behavioral intention and is, therefore, also
known as the goal-setting phase. Once the intention to
change one’s behavior has been developed, the volitional or
goal-pursuit phase begins.
The Health Action Process Approach (HAPA; Schwarzer,
Lippke, & Luszczynska, 2011) suggests phase-specific self-
efficacy beliefs. Motivational self-efficacy (e.g., “I am cer-
tain that I can do X”) is crucial to develop an intention in the
motivational phase. Volitional self-efficacy (e.g., “I am
capable of coping with barriers and recovering from set-
backs”), however, should be instrumental in pursuing goals
once they have been set (Schwarzer et al., 2011). It can com-
prise coping (or maintenance) and recovery self-efficacy,
and individuals already intending to change their behavior
should mainly benefit from such kinds of volitional self-
efficacy. Furthermore, action planning is a facilitating factor
in the volitional phase. It is a prospective self-regulatory
strategy, and individuals who generate plans are more likely
to translate their intentions into behavior (Koring et al.,
2012; Sniehotta, Scholz, & Schwarzer, 2005). However, the
belief in being able to change one’s behavior as well as
action planning is not sufficient. Earlier experience of con-
sequences that came along with this particular health behav-
ior (e.g., feeling better afterward) is also meaningful in the
behavior change process but the concept is not explicitly
included within the HAPA, which examines outcome expec-
tancies. The latter are prospective, whereas outcome experi-
ences relate to the same phenomenon, however in a
retrospective way.
The experience of behavioral consequences has to be dis-
tinguished from the concept of mastery experience as defined
by Bandura (1998). Mastery experience is seen as success
with the performance of the behavior itself and not as a pre-
viously experienced positive outcome of this behavior.
Fuchs, Goehner, and Seelig (2011) describe outcome experi-
ence as individual experience and appraisals after the adop-
tion of a novel behavior. The authors propose that past
experience of positive consequences of physical exercise
(e.g., “jogging made me feel better”) affects motivational
processes such as the formation of intentions. Based on this
assumption, a theory-based intervention program for inpa-
tients of an orthopedic rehabilitation clinic was developed
with prompting focus on past success as a behavior change
technique. More precisely, patients were asked to acknowl-
edge past positive consequences of physical exercise in a
group discussion session to set the motivational phase in
motion (Fuchs et al., 2011).
Furthermore, Rothman (2000) argues that being satisfied
with experienced consequences of a behavior is important to
Downloaded from heb.sagepub.com at University of Waikato Library on May 25, 2014
sustain an initiated behavior. His assumptions were con-
firmed by a study on rehabilitation patients’ positive experi-
ence with exercise (Fleig, Lippke, Pomp, & Schwarzer,
2011). In this study, the authors found that recently experi-
enced positive consequences enhanced satisfaction with
exercise and action planning, which in turn fostered exercise
maintenance. Their findings point to a mediating effect of
satisfaction and action planning between patients’ positive
experience with exercise and their subsequent levels of exer-
cise. However, the authors did not test whether positive
experience also had an impact on patients’ levels of exercise
via an increase in intentions and self-efficacy. Yet self-effi-
cacy beliefs are suggested to be fostered because of positive
experience as well, implying a mediating effect of self-effi-
cacy between positive experience and physical exercise
(Sherwood & Jeffery, 2000). This assumption was confirmed
in recent studies investigating a sample of university stu-
dents (Parschau et al., 2013) and a sample of older adults
with chronic illnesses (Warner, Schüz, Knittle, Ziegelmann,
& Wurm, 2011).
To sum up, there is some initial empirical evidence that
exercise experience may be a catalyst for several social–cog-
nitive mechanisms. However, the question of how positive
experience exerts its influence on physical exercise remains.
Aims
This research aims to investigate the role of exercise experi-
ence as a starting point in the behavior change processes.
Positive experience with consequences of physical exercise
is assumed to stimulate motivational self-efficacy, behav-
ioral intentions, and volitional self-efficacy. Action planning
should be affected by experience indirectly via an increase in
motivational self-efficacy, intentions, and volitional self-
efficacy. Action planning would, in turn, predict changes in
physical exercise. To evaluate the validity of this theoretical
model, we will test it in two distinct samples: the first origi-
nating from a longitudinal online observation study in the
general population, the second from a clinical intervention
study in a rehabilitation context.
Study 1
Method
Participants and Procedure. Participants were recruited
through a scientific TV show which is broadcasted weekly in
Germany. Data collection started with the program about
New Year’s resolutions in January 2012. The website of this
TV show contained a link to the online study. At Time 1
(T1), 729 participants provided informed consent, filled in
the online questionnaire, and provided their e-mail addresses
for follow-up assessments. The second measurement point in
time (T2) was implemented 2 weeks later by inviting partici-
pants via e-mail to respond to the follow-up questionnaire.
Parschau et al. 3

Table 1.  Intercorrelations, Factor Loadings, Means, Standard Deviations, and Ranges for Positive Experience, Motivational Self-Efficacy,
Volitional Self-Efficacy, Behavioral Intention, Action Planning and Exercise for Study 1.

1 2 3 4 5 6 M (SD) Range
1. Positive experience T1 .62-.72 2.91 (0.55) 1-4
2. Motivational self-efficacy T1 .28*** .81-.84 3.12 (0.72) 1-4
3. Volitional self-efficacy T2 .34*** .54*** .59-.84 3.10 (0.65) 1-4
4. Intention T1 .31*** .48*** .45*** .42-.86 3.45 (0.59) 1-4
5. Action planning T2 .27*** .38*** .40*** .47*** .51-.91 2.84 (0.84) 1-4
6. Exercise T1a .14* .36*** .27*** .31*** .29*** 102.66 (109.51) 0-487
7. Exercise T3a .18** .38*** .31*** .39*** .32*** .73*** 120.80 (120.48) 0-458

Note. All scores are related to the manifest scales. The ranges of indicator factor loadings on latent constructs are presented in the diagonal and in
boldface. Intercorrelations are presented below the diagonal. N = 350. T1 = Time 1; T2 = Time 2; T3 = Time 3.
a
Minutes per week.
*p < .05. **p < .01. ***p < .001.

The T2 questionnaire was completed by 555 participants.
Another e-mail invitation for the second follow-up question-
naire (T3) took place 5 weeks after T1. All three online ques-
tionnaires were completed by 350 participants.
The sample consisted of 63.7% women. Participants were
on average 41 years of age (SD = 12.8, range 16-90 years)
and had a mean body mass index of 24.7 kg/m2 (SD = 4.5,
range 16.8-49.5 kg/m2). The majority of the participants
were living with a partner (67.4%) and graduated from high
school (81.4%).
Measures.  The online-based questionnaires contained demo-
graphic and psychometric scales. All items given below were
translated from German.
Self-reported strenuous physical exercise was measured
twice by using two items of the Godin Leisure-Time Exercise
Questionnaire (Godin & Shephard, 1985; Plotnikoff et al.,
2007). At T1 and T3, participants were asked to indicate the
average number of sessions per week and the average dura-
tion of a session regarding strenuous physical exercise such
as intensive swimming, jogging, and cycling. Frequency and
average duration per session were multiplied to form a mea-
sure of weighted duration of strenuous exercise throughout a
normal week. Exercise scores 3 standard deviations larger
than the mean were truncated.
Positive exercise experience was measured at T1 by nine
items of the revised Exercise Experience Scale (Fleig et al.,
2011). The items were responded to on 4-point Likert-type
scales, ranging from 1 = completely disagree to 4 = com-
pletely agree. The item stem “When I was physically active,
I experienced that . . .” was followed by positive conse-
quences such as “. . . I felt physically better afterward”
(Cronbach’s α = .76). To minimize the number of single indi-
cators for the latent variable and thereby reducing error vari-
ances simultaneously, three parcels of three items each were
used as indicators for exercise experience (Little,
Cunningham, Shahar, & Widaman, 2002).
The following scales were adapted from Schwarzer et al.
(2011). The response format for each scale was a 4-point
Likert-type scale, ranging from 1 = completely disagree to 4
= completely agree and the time frame referred to “the next
weeks.”
Behavioral intention to perform physical exercise was
assessed at T1 with the item stem “I strongly intend to . . .”
which was followed by four items such as “. . . be physically
active several days a week” (Cronbach’s α = .77).
Action planning was assessed at T2 by three items
(Cronbach’s α = .81). The item stem “I have planned in detail
. . .” was followed by items such as “. . . how often I will be
physically active.”
Motivational self-efficacy consisted of two single-item
indicators assessed at T1 (r = .68) such as “I am certain that
I can be physically active even if it is difficult for me.”
Volitional self-efficacy was measured at T2 with four items
such as “I am certain that I can be physically active on a regular
basis even if I have to overcome myself” (Cronbach’s α = .83).
Intercorrelations, factor loadings, means, standard devia-
tions, and ranges of all constructs are displayed in Table 1.
Data Analysis.  SPSS 20 was used for reliability, descriptive
and attrition analyses. Structural equation modeling was per-
formed with AMOS 20. By default, missing values were
treated using full information maximum likelihood. Good-
ness-of-fit indices to evaluate model fit were the Tucker–
Lewis Index (TLI), comparative fit index (CFI), and root
mean square error of approximation (RMSEA). A satisfac-
tory model fit is indicated by TLI and CFI greater than .90
and RMSEA lower than .08 (Kline, 2005).
The structural equation model included positive exercise
experience as a distal predictor. Direct paths of positive
experience on motivational self-efficacy, intention, voli-
tional self-efficacy, and subsequent physical exercise were
added. Physical exercise at T1 was included as a covariate.
Furthermore, baseline exercise and positive experience were
allowed to correlate. Latent constructs were created for posi-
tive experience, motivational self-efficacy, behavioral inten-
tion, volitional self-efficacy, and action planning whereas
exercise was specified as a manifest outcome.

Downloaded from heb.sagepub.com at University of Waikato Library on May 25, 2014

4 Health Education & Behavior 
Figure 1.  Path diagram of Study 1 (N = 350).
Note. T1 = Time 1 (baseline); T2 = Time 2 (2 weeks after T1); T3 = Time 3 (5 weeks after T1). Nonsignificant paths are not displayed.
**p < .001.
Results
Attrition Analysis.  Attrition analyses revealed no differences
with regard to demographic characteristics (sex, age, partner
status, and education) and baseline assessments of positive
experience, motivational self-efficacy, and intention between
participants who provided complete longitudinal data (48.1%
of the initial sample) and those who did not take part in the
follow-up assessment. However, the longitudinal sample
was significantly more active at baseline (Mresponders = 104.21
minutes, SDresponders = 115.52, Mnonresponders = 84.56 minutes,
SDnonresponders = 114.29, t(727) = −2.31, p < .05). This differ-
ence was accounted for by including baseline physical exer-
cise as a covariate into the structural equation model.
Evaluation of the Sequential Mediator Model. The hypothe-
sized model had an adequate fit with χ2(126) = 325.79, p <
.001, χ2/df = 2.59, TLI = .90, CFI = .93 and RMSEA = .07
(90% confidence interval = [.06, .08]). Figure 1 displays its
standardized parameter estimates.
Positive experience emerged as a predictor of motiva-
tional self-efficacy and behavioral intention. Positive experi-
ence accounted for 16% of the variance in motivational
self-efficacy. The amount of the explained variance in inten-
tion was 41%.
The direct path from experience to physical exercise T3
was not significant (β = .07, p > .10). Therefore, the path was
omitted in Figure 1.
Downloaded from heb.sagepub.com at University of Waikato Library on May 25, 2014
Positive experience and motivational self-efficacy
accounted for 49% of the variance in volitional self-efficacy.
The amount of explained variance in physical exercise was
54%. Within the 5 weeks between T1 and T3, exercise levels
remained relatively stable.
Discussion
In an online-recruited sample, we investigated in which
way positive experience is associated with determinants of
the motivational and volitional phases of physical exer-
cise. Structural equation modeling provided evidence for
the relationship between experience and motivational self-
efficacy on one hand as well as behavioral intentions on
the other. Positive experience was also directly related to
volitional self-efficacy, even though not as strongly as to
motivational self-efficacy. Additionally, positive experi-
ence was not directly associated with changes in strenuous
exercise.
However, by interpreting the results, a selection bias
might have to be taken into account. It cannot be ruled out
that the volunteer sample might have had an enhanced inter-
est in the topic of physical exercise compared with the gen-
eral population. Therefore, it might be that the rather high
means in positive experiences, intention, and behavior are
associated with an enhanced interest in the study’s topic.
Furthermore, the investigated sample is likely to be biased
Parschau et al. 5
toward higher educated individuals and should therefore not
be generalized to individuals with a less educational
background.
Although data of three measurement points in time were
available, the relationship between positive experience,
motivational self-efficacy, and intention, respectively, was
examined only cross-sectionally at T1. An investigation of
four measurement points would allow testing the assumed
temporal order and the preceding role of exercise experience
in more detail. Therefore, we also analyzed data of a second
study with four measurement points in time.
Study 2
Method
Participants and Procedure.  Data of a second study were ana-
lyzed to substantiate the previous findings in the context of
tertiary prevention over four measurement points in time.
Participants were recruited in three German rehabilitation
clinics between 2009 and 2011 (one inpatient cardiac, one
inpatient orthopedic, and one outpatient orthopedic). The
regular program within these clinics provided medical, phys-
iotherapeutic, and psychological treatment. The first mea-
surement point in time (T1) took place at the beginning of
rehabilitation. After having provided informed consent, 461
patients completed a computer-based questionnaire. At the
end of the rehabilitation program (T2), patients were asked
to respond to a second computer-based questionnaire. Six
weeks after discharge, a third assessment was completed via
computer-assisted telephone interviews (CATI; T3). Six
months after discharge, the fourth assessment (T4) was again
carried out via CATI.
The longitudinal rehabilitation sample included partici-
pants who took part at least at the T1 and T4 measurements
(n = 275). Of this sample, 69% were individuals in orthope-
dic rehabilitation and 31% were individuals in cardiac reha-
bilitation. The sample consisted of more women (56%) than
men. The mean age was 50 years (SD = 9.3, range 19-76
years) and the mean body mass index was 28.2 kg/m2 (SD =
5.8, range 16.7-49.2 kg/m2). Most of the rehabilitation
patients indicated to have a partner (75.3%) and 44.7% of the
participants graduated from high school.
Measures.  The assessments contained demographic as well
as several psychometric scales. Social-cognitive variables
were partly related to the rehabilitation context and, there-
fore, differed from the ones in Study 1. All items given below
were translated from German.
Self-reported moderate and strenuous physical exercise
was measured by using four items of the modified version
of the Godin Leisure-Time Exercise Questionnaire (Godin
& Shephard, 1985; Plotnikoff et al., 2007). At T1 and T4,
participants were asked to indicate how many times and
how long per session they performed moderate (hardly
Downloaded from heb.sagepub.com at University of Waikato Library on May 25, 2014
exhausting, light sweating) and strenuous physical exercise
(fast heart rate, sweating) on average per week. For the
analyses, a composite score (number of sessions per week
multiplied by minutes per session) was formed for moder-
ate and strenuous activities. Both scores (r = .29) were then
aggregated to the total number of exercise per week. Scores
larger than three standard deviations above the sample’s
mean were truncated.
Based on a measure by Lippke, Fleig, Pomp, and
Schwarzer (2010), behavioral intentions to perform exer-
cise were assessed at T2 with the item stem “I intend to do
the following activities . . .” which was followed for exam-
ple by “. . . fitness activities (e.g., using an exercise bike).”
Participants could indicate the frequency and duration they
intended by choosing one of the following answering
options: 1 = not at all, 2 = less than one time per week for
about 40 [20] minutes, 3 = at least once per week for 40
[20] minutes, 4 = at least three times per week for 40 [20]
minutes, and 5 = at least five times per week for 40 [20]
minutes or more. Because of different medical recommen-
dations, response options regarding the duration varied for
orthopedic (20 minutes) and cardiac (40 minutes) rehabili-
tation patients. To account for this difference, the type of
patient was entered as a covariate into the structural equa-
tion models. Even though the internal consistency for this
scale was low (Cronbach’s α = .44), a factor analysis sug-
gested one component for all three items (factor loadings
between .56 and .80).
The response format for all following items was a 6-point
Likert-type scale, ranging from 1 = completely disagree to 6
= completely agree.
Positive exercise experience was measured at T1 by five
items of the revised Exercise Experience Scale (Fleig et al.,
2011). The item stem “When I was physically active, I expe-
rienced that . . .” was followed by positive consequences
such as “. . . it had a positive impact on my health” (Cronbach’s
α = .75).
Action planning was assessed at T3 by four items
(Cronbach’s α = .87). The item stem “Regarding the next
four weeks, I have already planned . . .” was followed for
example by the item “. . . which physical activities I will
perform.”
Phase-specific self-efficacy was assessed at T2 (Lippke
et al., 2010). Motivational self-efficacy was measured with
the stem “I am certain that I can be physically active on a
regular basis even if . . .” followed by two items such as “. . .
it is difficult” (r = .83). Volitional self-efficacy consisted of
two indicators such as “I am capable of exercising on a regu-
lar basis even if it takes some time until it becomes a routine”
(r = .67). The time frame of the two self-efficacy scales
referred to “the next weeks after the rehabilitation.”
Intercorrelations, factor loadings, means, standard devia-
tions, and ranges for each manifest scale are displayed in
Table 2. All analyses were performed in the same manner as
in Study 1.
6 Health Education & Behavior 

Table 2.  Intercorrelations, Factor Loadings, Means, Standard Deviations, and Ranges for Positive Experience, Motivational Self-Efficacy,
Volitional Self-Efficacy, Behavioral Intention, Action Planning, and Exercise for Study 2.

1 2 3 4 5 6 M (SD) Range
1. Positive experience T1 .41-.81 3.96 (1.11) 1-6
2. Motivational self-efficacy T2 .21** .90-.92 4.80 (1.24) 1-6
3. Volitional self-efficacy T2 .04 .53*** .68-.99 4.15 (1.41) 1-6
4. Intention T2 .31*** .16* .09 .32-.57 2.88 (0.72) 1-5
5. Action planning T3 .12 .27*** .24*** .24*** .74-.88 4.46 (1.45) 1-6
6. Exercise T1a .16* .20** .10 .20** .14* 75.25 (99.26) 0-354
7. Exercise T4a .15* .22** .04 .30*** .20** .29*** 149.82 (121.87) 0-464

Note. All scores are related to the manifest scales. The ranges of indicator factor loadings on latent constructs are presented in the diagonal and in
boldface. Intercorrelations are presented below the diagonal. N = 275. T1 = Time 1; T2 = Time 2; T3 = Time 3; T4 = Time 4.
a
Minutes per week.
*p < .05.**p < .01. ***p < .001.

Results Positive experience at the beginning of the rehabilitation pro-

Attrition Analysis. At baseline, 461 rehabilitation patients
agreed to participate in the study. The first follow-up ques-
tionnaire (T2) was completed by 377 participants (81.8%) at
the end of rehabilitation. At T3, 346 of them responded to the
second follow-up questionnaire (CATI, 75.1% of the initial
sample) and at T4, 275 responded to the third follow-up
questionnaire (CATI, 59.7% of the initial sample).
Attrition analyses revealed no baseline differences in
terms of sex, type of patients, partner status, education, exer-
cise, and positive experience between participants who
stayed in the longitudinal sample and those who did not take
part in the follow-up assessments. The longitudinal sample
was, however, older (Mresponders = 50.2, SDresponders = 9.3,
Mnonresponders = 47.8, SDnonresponders = 11.6, t(314.6) = −2.24,
p < .05; effect size d = .17). Therefore, age was included as a
covariate into the structural equation model.
Evaluation of the Sequential Mediator Model. The fit of the
hypothesized model was satisfactory with χ2(161) = 278.60,
p < .001, χ2/df = 1.73, TLI = .90, CFI = .93, and RMSEA =
.05 (90% confidence interval = [.04, .06]). Figure 2 displays
the standardized parameter estimates.
Levels of motivational self-efficacy and behavioral inten-
tions were predicted by levels of positive experience. Positive
experience accounted for 6% of the variance in motivational
self-efficacy. The amount of explained variance in intentions
was 28%.
Neither volitional self-efficacy (β = −.06, p > .10) nor
exercise at T4 (β = .12, p = .10) were significantly predicted
by positive experience. The amount of explained variance in
exercise was 13%.
No significant effects of type of patient (orthopedic vs.
cardiac) and age were found on exercise levels.
gram was associated with motivational self-efficacy as well
as with behavioral intentions at the end of the rehabilitation.
However, positive experience at the beginning of the rehabili-
tation was neither related to volitional self-efficacy 6 weeks
after discharge nor related to physical exercise 6 months after
discharge.
Although the setting and sample characteristics differed
between both studies, all paths of the hypothesized model—
apart from the path between exercise experience and voli-
tional self-efficacy—were replicated in Study 2.
General Discussion
As an individual’s current exercise may partly be driven by
past experiences with exercise consequences, we investigated
the function of positive exercise experience in two distinct
samples. Changes in the level of physical exercise were not
directly predicted by positive experience neither in the online-
recruited sample of the general population nor in the sample
of rehabilitation patients. Positive exercise experience did,
however, relate to motivational self-efficacy and behavioral
intentions, which in turn predicted later exercise. Additionally,
positive experience was associated with volitional self-effi-
cacy in Study 1. Nevertheless, results of both studies indicate
a more consistent link between positive exercise experience
and motivational factors, less so with volitional factors. This
finding is in accordance with the assumptions by Fuchs et al.
(2011) suggesting that the experience of positive conse-
quences affects motivational determinants of exercise.
Furthermore, both studies found intentions as well as voli-
tional self-efficacy to be related to subsequent action planning
which, in turn, was associated with exercise later on. These
findings partly confirm previous evidence for the HAPA
(Chiu, Lynch, Chan, & Berven, 2011; Sniehotta et al., 2005).

Discussion Strengths and Limitations

In Study 2, the sample consisted of rehabilitation patients, A strength of this article lies in the replication of findings
and data were assessed at four measurement points in time. because the hypothesized model was tested in two samples

Downloaded from heb.sagepub.com at University of Waikato Library on May 25, 2014

Parschau et al. 7
Figure 2.  Path diagram of Study 2 (N = 275).
Note. T1 = Time 1 (baseline); T2 = Time 2 (end of rehabilitation); T3 = Time 3 (6 weeks after T2); T4 = Time 4 (6 months after T2). Nonsignificant paths
are not displayed.
*p < .05. **p < .001.
and distinct contexts. Although both studies showed a con-
siderable dropout rate, the samples had an acceptable size
with 350 and 275 study participants, respectively. However,
the differences between responders and nonresponders were
accounted for in our analyses. Although the samples were
surveyed on three and four measurement points in time in
Study 1 and Study 2, respectively, causality cannot be proven
by these research designs.
Implications for Future Research
Previous studies found motivational self-efficacy to be a
strong predictor of behavioral intentions (Parschau et al.,
2012; Scholz, Sniehotta, & Schwarzer, 2005). However, evi-
dence on how motivational self-efficacy can be fostered is
rare. On the basis of the literature (Fleig et al., 2011; Fuchs
et al., 2011; Parschau et al., 2013; Rothman, 2000) and the
results of the present study, the construct of exercise experi-
ence is a promising candidate to be integrated and further
investigated in health behavior change models.
As both studies investigated a selection of motivational
and volitional components taken from established models
only (e.g., HAPA), the function and location of exercise
experience in health behavior change models is not yet suf-
ficiently clarified. In future studies, direct and indirect
associations between positive experience and volitional
Downloaded from heb.sagepub.com at University of Waikato Library on May 25, 2014
cognitions such as action planning, coping planning, and
action control might be investigated in more detail by dif-
ferentiating between the initiation and maintenance of
physical exercise. Moreover, examining whether long-term
or short-term exercise experience has different effects on
subsequent exercise might be of further interest. The the-
matic content of the experience (e.g., health, emotion, or
appearance related) might also offer implications for exer-
cise interventions.
These two studies bear several practical implications for
the promotion of exercise self-efficacy, exercise intentions,
and exercise behavior in the general population and in reha-
bilitation patients. To raise individuals’ awareness of positive
exercise experience, a recall of positive effects of previous
exercise could be prompted in interviews or recorded in dia-
ries (Fleig et al., 2011; Pomp, Fleig, Schwarzer, & Lippke,
2012). Providing individuals with feedback on their positive
exercise outcomes has been found effective in previous stud-
ies (Ashford, Edmunds, & French, 2010) and could be
another straightforward strategy to be implemented, for
example, in the rehabilitation context.
Conclusion
According to the present results, positive exercise experience
seems to be relevant in the motivational phase of changes in
8 Health Education & Behavior 
exercise behavior. Together with future research on positive
experience in the context of other health behaviors, this con-
struct is a promising candidate to be integrated into the moti-
vational phase of health behavior process models.
Acknowledgments
We thank the rehabilitation clinics and their rehabilitants for par-
ticipating in this study. We especially appreciate the support of Mrs.
Pimmer, Dr. Kiwus, Dr. Glatz, Dr. Milse, and Dr. Johnigk.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interests with respect
to the research, authorship, and/or publication of this article.
Funding
The authors declared the following financial support for the research,
authorship, and/or publication of this article: Study 2 was supported
by the Deutsche Rentenversicherung Bund (DRV; German Pension
Insurance) within the project FABA (Project ID 8011-106-31/31.91).
References
Aaltonen, S., Leskinen, T., Morris, T., Alen, M., Kaprio, J.,
Liukkonen, J., & Kujala, U. (2012). Motives for and barriers
to physical activity in twin pairs discordant for leisure time
physical activity for 30 years. International Journal of Sports
Medicine, 33, 157-163. doi:10.1055/s-0031-1287848
Ashford, S., Edmunds, J., & French, D. P. (2010). What is the
best way to change self-efficacy to promote lifestyle and rec-
reational physical activity? A systematic review with meta-
analysis. British Journal of Health Psychology, 15, 265-288.
doi:10.1348/135910709X461752
Bandura, A. (1998). Health promotion from the perspective of
social cognitive theory. Psychology & Health, 13, 623-649.
doi:10.1080/08870449808407422
Chiu, C. Y., Lynch, R. T., Chan, F., & Berven, N. L. (2011). The
Health Action Process Approach as a motivational model for
physical activity self-management for people with multiple
sclerosis: A path analysis. Rehabilitation Psychology, 56,
171-181. doi:10.1037/a0024583
Fleig, L., Lippke, S., Pomp, S., & Schwarzer, R. (2011). Exercise
maintenance after rehabilitation: How experience can make
a difference. Psychology of Sport and Exercise, 12, 293-299.
doi:10.1016/j.psychsport.2011.01.003
Fuchs, R., Goehner, W., & Seelig, H. (2011). Long-term effects
of a psychological group intervention on physical exercise and
health: The MoVo concept. Journal of Physical Activity and
Health, 8, 794-803.
Godin, G., & Shephard, R. J. (1985). A simple method to assess
exercise behavior in the community. Canadian Journal of
Applied Sport Sciences, 10, 141-146.
Heckhausen, H. (1991). Motivation and action. Berlin, Germany:
Springer.
Kline, R. B. (2005). Principles and practice of structural equation
modeling. New York, NY: Guilford Press.
Koring, M., Richert, J., Lippke, S., Parschau, L., Reuter, T.,
& Schwarzer, R. (2012). Synergistic effects of planning
Downloaded from heb.sagepub.com at University of Waikato Library on May 25, 2014
and self-efficacy on physical activity. Health Education &
Behavior, 39, 152-158. doi:10.1177/1090198111417621
Lippke, S., Fleig, L., Pomp, S., & Schwarzer, R. (2010). Validity
of a stage algorithm for physical activity in participants
recruited from orthopedic and cardiac rehabilitation clin-
ics. Rehabilitation Psychology, 55, 398-408. doi:10.1037/a00
21563
Little, T. D., Cunningham, W. A., Shahar, G., & Widaman,
K. F. (2002). To parcel or not to parcel: Exploring the question,
weighing the merits. Structural Equation Modeling, 9,
151-173. doi:10.1207/S15328007SEM0902_1
Parschau, L., Fleig, L., Koring, M., Lange, D., Knoll, N., Schwarzer,
R., & Lippke, S. (2013). Positive experience, self-efficacy, and
action control predict physical activity changes: A moderated
mediation analysis. British Journal of Health Psychology, 18,
395-406. doi:10.1111/j.2044-8287.2012.02099.x
Parschau, L., Richert, J., Koring, M., Ernsting, A., Lippke, S., &
Schwarzer, R. (2012). Changes in social-cognitive variables
are associated with stage transitions in physical activity. Health
Education Research, 27, 129-140. doi:10.1093/her/cyr085
Plotnikoff, R. C., Lippke, S., Reinbold-Matthews, M., Courneya,
K. S., Karunamuni, N., Sigal, R. J., & Birkett, N. (2007).
Assessing the validity of a stage measure on physical activity
in a population-based sample of individuals with type 1 or type
2 diabetes. Measurement in Physical Education and Exercise
Science, 11, 73-91. doi:10.1080/10913670701294062
Pomp, S., Fleig, L., Schwarzer, R., & Lippke, S. (2012). Depressive
symptoms interfere with post-rehabilitation exercise: Outcome
expectancies and experience as mediators. Psychology, Health,
& Medicine, 17, 698-708. doi:10.1080/13548506.2012.661864
Rothman, A. J. (2000). Toward a theory-based analysis of behav-
ioral maintenance. Health Psychology, 19(1 Suppl.), 64-69.
doi:10.1037/0278-6133.19.Suppl1.64
Saxena, S., Van Ommeren, M., Tang, K. C., & Armstrong, T. P.
(2005). Mental health benefits of physical activity. Journal
of Mental Health, 14, 445-451. doi:10.1080/09638230500
270776
Scholz, U., Sniehotta, F. F., & Schwarzer, R. (2005). Predicting
physical exercise in cardiac rehabilitation: The role of phase-
specific self-efficacy beliefs. Journal of Sport & Exercise
Psychology, 27, 135-151.
Schwarzer, R., Lippke, S., & Luszczynska, A. (2011). Mechanisms
of health behavior change in persons with chronic illness or
disability: The Health Action Process Approach (HAPA).
Rehabilitation Psychology, 56, 161-170. doi:10.1037/a00
24509
Sherwood, N. E., & Jeffery, R. W. (2000). The behavioral deter-
minants of exercise: Implications for physical activity inter-
ventions. Annual Review of Nutrition, 20, 21-44. doi:10.1146/
annurev.nutr.20.1.21
Sniehotta, F. F., Scholz, U., & Schwarzer, R. (2005). Bridging the inten-
tion-behaviour gap: Planning, self-efficacy, and action control in
the adoption and maintenance of physical exercise. Psychology &
Health, 20, 143-160. doi:10.1080/08870440512331317670
Taylor, W. C., Yancey, A. K., Leslie, J., Murray, N. G., Cummings,
S. S., Sharkey, S. A., . . .McCarthy, W. J. (1999). Physical activ-
ity among African American and Latino middle school girls:
Consistent beliefs, expectations, and experiences across two sites.
Women & Health, 30(2), 67-82. doi:10.1300/J013v30n02_05
Parschau et al. 9

Toscos, T., Consolvo, S., & McDonald, D. W. (2011). Barriers
to physical activity: A study of self-revelation in an online
community. Journal of Medical Systems, 35, 1225-1242.
doi:10.1007/s10916-011-9721-2
Warburton, D. E., Nicol, C. W., & Bredin, S. S. (2006). Health
benefits of physical activity: The evidence. Canadian Medi­
cal Association Journal, 174, 801-809. doi:10.1503/cmaj.
051351
Warner, L. M., Schüz, B., Knittle, K., Ziegelmann, J. P., & Wurm, S.
(2011). Sources of perceived self-efficacy as predictors of physi-
cal activity in older adults. Applied Psychology: Health and
Well-Being, 3, 172-192. doi:10.1111/j.1758-0854.2011.01050.x
World Health Organization. (2011). Global status report on non-
communicable diseases 2010. Description of the global bur-
den of NCDs, their risk factors and determinants. Geneva,
Switzerland: Author.

Downloaded from heb.sagepub.com at University of Waikato Library on May 25, 2014