Effects of an intervention based on self-determination theory on self-reported leisure-time physical activity participation

Abstract

Based on self-determination theory, the present study developed and evaluated the utility a school-based intervention to change pupils' physical activity intentions and self-reported leisure-time physical activity behaviour. The study evaluated utility of the intervention to promote physical activity participation over a 5-week interval of time. A cluster randomised design targeting 215 pupils from 10 schools with schools as the unit of randomisation was adopted (Male = 106, Female = 109, Age = 14.84, SD = 0.48). Results indicated that pupils who were taught by autonomy-supportive teachers reported stronger intentions to exercise during leisure time and participated more frequently in leisure-time physical activities than pupils in the control condition. Autonomous motivation and intentions mediated the effects of the intervention on self-reported physical activity behaviour. It is concluded that self-determination theory provides a useful framework for the development of school-based interventions that ultimately affect leisure-time physical activity participation.

Full text

Planned behaviour and self-determination theory 1
Running head: PLANNED BEHAVIOUR AND SELF-DETERMINATION THEORY
1
2
3
4
Integrating the theory of planned behaviour and self-determination theory in health behaviour:
5
A meta-analysis
6
7
8
Martin S. Hagger
9
University of Nottingham
10
Nikos L. D. Chatzisarantis
11
National Institute of Education, Singapore
12
13
14
15
Author Note
16
Nikos L.D. Chatzisarantis, National Institute of Education, Nanyang Technological
17
University, 1 Nanyang Walk, Singapore 637616, email: nikos.chatzisarantis@nie.edu.sg
18
Correspondence concerning this chapter should be addressed to Martin S. Hagger,
19
School of Psychology, University of Nottingham, University Park, Nottingham, NG7 2RD,
20
United Kingdom, email: martin.hagger@nottingham.ac.uk
21
22

Planned behaviour and self-determination theory 2
Running head: PLANNED BEHAVIOUR AND SELF-DETERMINATION THEORY
1
2
3
4
Integrating the theory of planned behaviour and self-determination theory in health behaviour:
5
A meta-analysis
6
7
8
9

Planned behaviour and self-determination theory 3
Abstract
1
Purpose. A meta-analysis of studies integrating the theory of planned behaviour (TPB)
2
and self-determination theory (SDT) in health contexts is presented. The analysis aimed to
3
provide cumulative empirical support for a motivational sequence in which self-determined
4
motivation from SDT predicts the proximal predictors of intentions and behaviour from the
5
TPB.
6
Methods. A literature search identified 34 integrated studies providing 43 tests of effects
7
between TPB and SDT variables. Hunter and Schmidt’s (1994) methods of meta-analysis were
8
used to correct the effect sizes across the studies for statistical artifacts. Age (old vs. young),
9
publication status (published vs. unpublished), study design (correlational vs.
10
experimental/intervention), and behaviour type (physical activity vs. other health-related
11
behaviours) were evaluated as moderators of the effects. A path-analysis using the meta-
12
analytically derived correlations was conducted to examine the proposed motivational
13
sequence.
14
Results. Statistically-significant corrected correlations were evident among the perceived
15
autonomy support and self-determined motivation constructs from SDT and the attitude,
16
subjective norms, perceived behavioural control, intention, and health-related behaviour
17
constructs from the TPB. Only six of the 28 effect sizes were moderated by the proposed
18
moderators. Path analysis revealed that the significant effects of self-determined motivation on
19
intentions and behaviour were partially mediated by the proximal predictors from the TPB.
20
Conclusions. Evidence from this synthesis supported the theoretical integration and
21
proposed motivational sequence. Results are discussed with reference to the complimentary
22
aspects of the TPB and SDT and the need for integrated experimental or intervention studies on
23
a broader range of health behaviours.
24
25

Planned behaviour and self-determination theory 4
Research into the antecedent factors and processes that underpin people’s motivation to
1
engage in health-related behaviour has been conducted from an array of different theoretical
2
perspectives (Conner & Norman, 2005; Johnston, 2005; Orbell, 2004). Prominent among these
3
theories is the planned behaviour (TPB, Ajzen, 1985), which was developed as a systematic
4
explanation of volitional behavioural engagement based on a set of belief-based perceptions,
5
the deliberative formation of intentions, and their enactment. This approach has been shown to
6
account for substantial variance in behaviour in a number of contexts (Armitage & Conner,
7
2001; Conner & Armitage, 1998). In contrast, organismic approaches to motivation have
8
focused on the contextual contingencies and dispositional orientations that give rise to
9
motivational states and subsequent behaviour. One such approach is self-determination theory
10
(SDT, Deci & Ryan, 1985, 2000), a leading theory of human motivation that has been shown
11
to be efficacious in identifying the contingencies that affect motivation and behaviour in
12
numerous domains (Deci & Ryan, 1985, 2000).
13
While independent meta-analyses have shown both theories to be effective in accounting
14
for variation in health-related behaviour (e.g., Chatzisarantis, Hagger, Biddle, Smith, & Wang,
15
2003; Hagger, Chatzisarantis, & Biddle, 2002b; Sheeran & Orbell, 1998), both have
16
shortcomings in terms of their predictive utility. SDT does not chart the exact process by which
17
motivational orientations are converted into intentions and behaviour and the TPB has
18
provided an effective basis for the explanation of variance in intentions and health-related
19
behaviour without identifying the origins of the antecedents of the behaviour (Chatzisarantis,
20
Hagger, & Smith, 2007).
21
In this article we propose that these two prominent social psychological theories can
22
serve to assist in overcoming these shortcomings by integrating their constructs and hypotheses
23
in a unified model of motivation to explain intentions and health-related behaviour. The basis
24
for integration is offered by Deci and Ryan (1985) and Vallerand (1997) who state that
25
motivational theories can offer explanations for the origins of social cognitive beliefs and
26

Planned behaviour and self-determination theory 5
expectations outlined in models of intention like the TPB. It is proposed that individuals with a
1
self-determined motivational orientation towards a health-related behaviour will tend to form
2
attitudes and perceptions of control, two key determinants of intention from the TPB,
3
congruent with their motivational orientations (Hagger & Chatzisarantis, 2008b; Hagger et al.,
4
in press; Vallerand, 2007). Support for the proposed integrated model and motivational
5
sequence will be provided in a meta-analysis of the extant research on health-related
6
behaviours that has integrated the TPB and SDT.
7
Component Theories of the Integrated Model
8
Prior to outlining the rationale behind the theoretical integration, a brief explanation of
9
the hypotheses of each component theory is warranted. The TPB posits that an individual’s
10
intention is the most proximal predictor of health-related behaviour and mediates the effect of
11
three sets of belief-based perceptions on behaviour: attitudes, subjective norm, and perceived
12
behavioural control (PBC) (Ajzen, 1985). Attitudes are a person’s overall positive or negative
13
evaluation of the target behaviour. Subjective norms summarise a person’s expectation that
14
significant others want them to engage in the target behaviour. PBC comprises a person’s
15
overall judgment whether they have the ability and resources available to engage in the target
16
behaviour.
17
Self-determination theory (SDT) takes a different approach and focuses on the quality of
18
an individual’s motivation in a given context and the environmental factors that affect
19
motivation in that context (Deci & Ryan, 1985; Ryan & Connell, 1989). Central to the theory is
20
the distinction between self-determined versus controlled types of motivation (Deci & Ryan,
21
2000). Individuals whose motivation is self-determined experience a sense of personal choice
22
and autonomy when behaving and feel their actions represent their true self. Those whose
23
motivation is not self-determined feel controlled, pressured, or coerced into behaving by
24
external forces. SDT research has shown that self-determined motives positively affect
25

Planned behaviour and self-determination theory 6
behavioural engagement (Chatzisarantis et al., 2003) and that self-determined motivation can
1
be supported or thwarted by environmental contingencies, such as the support offered by
2
salient others (Hagger et al., 2007; Reeve, Bolt, & Cai, 1999).
3
Integrating the Theories
4
Recently, researchers have sought to integrate SDT and the TPB because these
5
approaches are deemed to provide complimentary explanations of the processes that underlie
6
motivated behaviour. (e.g., Ntoumanis, 2001; Sarrazin, Vallerand, Guillet, Pelletier, & Cury,
7
2002; Standage, Duda, & Ntoumanis, 2003; Wilson & Rodgers, 2004). Numerous authors have
8
proposed that organismic theories of motivation like SDT could potentially offer explanations
9
for the origins of constructs in social cognitive theories (Andersen, Chen, & Carter, 2000; Deci
10
& Ryan, 1985). The integration is based on the links between self-determined motivation and
11
the system of beliefs that underpin the proximal antecedents of intention: attitudes, subjective
12
norms, and PBC.
13
Ajzen (1985) proposed that attitudes and PBC comprise beliefs that a given health-related
14
behaviour will result in certain outcomes and that the behaviour is under control of the
15
individual. However, like many social cognitive theories, the TPB is not explicit in the reasons
16
that these beliefs are pursued (Deci & Ryan, 1985). For example, the theory does not make the
17
distinction between beliefs about outcomes that people choose to seek and are related to their
18
true sense of self (self-determined outcomes) and beliefs about outcomes that people feel
19
compelled to engage in out of a sense of obligation or duty (controlled outcomes). Some
20
beliefs about outcomes could be interpreted as either self-determined or controlled reasons for
21
participating in health-related behaviour (e.g., “I exercise in order to lose weight”). For some
22
people exercising to lose weight may be self-determined because they value being healthy and
23
view it as representative of their true self (self-determined). They are therefore likely to
24
actively seek opportunities to exercise to lose weight. For others losing weight may be to look
25

Planned behaviour and self-determination theory 7
good for others, an external contingency (controlling). In such cases, they are not likely to
1
pursue exercise to lose weight or may even avoid it. Therefore, SDT offers an interpretation of
2
whether these beliefs about outcomes are interpreted as self-determined or controlling. The
3
theory suggests that motivation to engage in health-related behaviours for self-determined or
4
controlling reasons predisposes individuals to form beliefs congruent with these motives.
5
On this basis, self-determined motives are hypothesised to be a distal predictor of
6
attitudes and PBC. Attitudes and PBC are, in turn, proximal predictors of the formation of
7
intentions to engage in future health-related behaviour in accordance with the TPB. Therefore a
8
motivational sequence is proposed such that the effects of self-determined motivation on
9
intentions and health-related behaviour are mediated by the proximal determinants, namely,
10
attitudes and PBC (Hagger, Chatzisarantis, & Harris, 2006).
11
The Present Study
12
The literature testing the proposed relations between variables from the TPB and SDT in
13
health contexts is substantial. We aimed to test whether there was consistency in the pattern of
14
the proposed relationships in the motivational sequence across these studies using meta-
15
analytic techniques. This is important as it will provide support for the proposed sequence
16
based on the available literature and resolve any inconsistencies attributable to methodological
17
artefacts such as sampling and measurement error.
18
We also propose to test the effectiveness of five key moderator variables in accounting
19
for additional variation in the effect sizes among the integrated model constructs: age, gender,
20
publication status, study design, and behaviour type. Previous meta-analyses have examined
21
these moderator variables in syntheses of research on the TPB and SDT separately
22
(Chatzisarantis et al., 2003; Hagger et al., 2002b; Sheeran & Orbell, 1998). Age and gender
23
will be examined as moderators to test the hypothesis that the proposed effects are universal.
24
Evidence for publication bias in the hypothesised effects among constructs in the current
25

Planned behaviour and self-determination theory 8
integrated theoretical approach will be evaluated by including publication status as a
1
moderator. Study design was tested as a moderator to ensure that there was no variation in the
2
effects due to the adoption of an experimental or correlational design. Finally, the universality
3
hypothesis suggests that the proposed effects should be invariant across health behaviours and
4
a moderator analysis by behaviour type seeks to provide evidence to support this.
5
In addition, the proposed motivational sequence can be tested empirically using the meta-
6
analytically derived effect sizes in a path analysis (Viswesvaran & Ones, 1995). This provides
7
corroborative evidence for the unique effects in the motivational sequence and also tests for
8
key mediation effects. Specifically, a meta-analytic path analysis model will be tested with
9
self-determined motivation from SDT set to predict the proximal predictors of intentions to
10
engage in health-related behaviour from the TPB, namely, attitudes, subjective norms, and
11
PBC. It is expected that self-determined motivation be strongly related to attitudes and PBC as
12
in previous studies (Chatzisarantis, Hagger, Biddle, & Karageorghis, 2002; Hagger,
13
Chatzisarantis, & Biddle, 2002a). However, self-determined motivation is not expected to be
14
related to subjective norms because the latter is defined as controlling rather than self-
15
determined perceptions and this is supported by tests of this relationship have yielded small or
16
negative effects (Hagger, Chatzisarantis, Barkoukis, Wang, & Baranowski, 2005; Hagger,
17
Chatzisarantis, Culverhouse, & Biddle, 2003). Further, attitudes, subjective norms, and PBC
18
are expected to predict intentions in keeping with proposals from the TPB (Ajzen, 1985).
19
Attitudes and PBC are expected to mediate the effect of self-determined motivation on
20
intentions. Finally, intentions are proposed to mediate the effects of attitudes, subjective norms,
21
and PBC on health-related behaviour, in accordance with the TPB. The sequence is expected to
22
mediate the effects of self-determined motivation on health behaviour such that there is no
23
direct effect of self-determined motivation on behaviour (e.g., Hagger et al., 2003).
24

Planned behaviour and self-determination theory 9
In addition, numerous studies integrating the TPB and SDT have included perceived
1
autonomy support as a variable that reflects whether the environment is perceived to support
2
self-determined motivation (e.g., Hagger et al., in press; Shen, McCaughtry, & Martin, 2007).
3
Therefore, perceived autonomy support will also be included as an independent predictor of
4
self-determined motivation. Finally, the proposed effects are expected to be independent of
5
past behaviour and we will control for past behaviour in the meta-analytic path analysis.
6
Method
7
Literature search
8
Research articles were located via an exhaustive search of electronic databases (e.g.,
9
Medline, PsychINFO, Psyarticles, ISI Web of Science), manual searches of key journals, and
10
the reference sections of review articles on SDT and the TPB to the end of July 20081. The aim
11
of the literature search was to establish a fully-inclusive database of integrated tests of effects
12
between constructs from the TPB and SDT and to ensure that all reasonable attempts had been
13
made to locate ‘fugitive literature’ (Rosenthal, 1994).
14
Inclusion criteria
15
Studies were included if they provided an empirical test of an effect size from the
16
integration of the TPB and SDT in health-related behaviour contexts. Studies therefore had to
17
include effect sizes from a measure of at least one construct from SDT and one from the TPB
18
or, its predecessor, the theory of reasoned action (Ajzen & Fishbein, 1980). In addition, studies
19
were required to contain sufficient statistical information such as zero-order correlation
20
coefficients in correlational studies or cell means, standard deviations, F-ratios, or effect size
21
statistics (e.g., Cohen’s d) in experimental or intervention studies to calculate an effect size.
22
One of the challenges of meta-analytic syntheses of studies is establishing equivalence
23
among the measures used in tests of the desired effect size. In the present analysis, only a small
24

Planned behaviour and self-determination theory 10
subset of measures was used to tap the TPB and self-determination theory constructs across
1
studies. This made identifying studies that had tested specific effects relatively unambiguous.
2
The TPB constructs were exclusively measured using items derived directly or indirectly from
3
the guidelines proposed by Ajzen (2003). SDT constructs were invariably derived from
4
variants of the perceived locus of causality scales or the relative autonomy index or self-
5
determination index, which are reweighted aggregates of the self-determined and non-self-
6
determined motivational constructs from the perceived locus of causality (Ryan & Connell,
7
1989)2. Experimental or intervention studies based on the theories used conventional methods
8
to manipulate key variables such as self-determined motivation using autonomy support (e.g.,
9
Reeve et al., 1999).
10
The literature search yielded 34 studies that met the search criteria. The majority of the
11
studies were correlational in design with a small proportion reporting experimental or
12
intervention designs based on SDT and reporting effects on TPB constructs like intention (e.g.,
13
Chatzisarantis et al., 2007, Study 3; Edmunds, Ntoumanis, & Duda, in press). There were no
14
studies that had manipulated TPB variables and examined the effects on SDT constructs.
15
Studies that integrated the two theories and but did not target a health-related behaviour were
16
excluded (e.g., Lin, 2007; Phillips, Abraham, & Bond, 2003). There were 43 independent tests
17
of effect sizes among the constructs from both theories. The highest number of tests of an
18
individual effect was 38 for the association between self-determined motivation and intention.
19
Meta-analytic strategy
20
We used Hunter and Schmidt’s (1994) methods for meta-analysis to correct the effect
21
sizes for statistical artifacts. Statistical theorists have advocated the adoption of such random
22
effects models for meta-analysis as these are optimal in permitting the generalization of
23
corrected effect sizes to the population (Field, 2001; Hagger, 2006; Hunter & Schmidt, 2000).
24
We corrected for two statistical artifacts in the present study: sampling error and measurement
25

Planned behaviour and self-determination theory 11
error. The effect size of choice was the zero-order correlation coefficient as it was the most
1
frequently adopted metric. Studies reporting other metrics were converted into correlation
2
coefficients using Hunter and Schmidt’s (1994) algorithms. We corrected for measurement
3
error using the reliability statistics, usually Cronbach alpha coefficients, of the constructs used
4
in each effect size calculation. Where reliability statistics were unavailable, measurement error
5
was inferred from available attenuation statistics using a formula supplied by Stauffer (1996).
6
The analysis produces a ‘bare bones’ correlation coefficient (r+) representing the mean
7
average effect size corrected for sampling error only and a fully-corrected correlation
8
coefficient (r++) which is the mean average effect size corrected for both sampling and
9
measurement error. 95% confidence intervals (CI95) reflecting the distribution about the mean
10
effect size are used to test the statistical significance of the average corrected effect size. If the
11
CI95 does not include the value of zero, then it is likely that the effect exists in the population
12
(Hunter & Schmidt, 1994). The analysis also yields 90% credibility intervals (CI90)
13
representing the distribution about the average corrected correlation using the residual standard
14
deviation and provides an estimation of the distribution of the effect size in the population.
15
This is used to evaluate the discriminant validity of the constructs i.e. the hypothesis that
16
population effect size is significantly different from unity. We also computed the ‘fail safe’
17
sample size (NFS) which represents the number of studies with null findings required reduce the
18
effect size to a trivial value, in this case resulting in the confidence intervals including the
19
value of zero (Rosenberg, 2005). If the number of ‘null finding’ tests of an effect is sufficiently
20
large the researcher can be confident that the chances of such a number of studies existing is
21
improbable. Finally, the percentage variance in the effect sizes across studies attributed to the
22
corrected artifacts relative to the overall variance in the effect size is given. Hunter and
23
Schmidt (1994) advocate a 75% cutoff criterion for an effect size to be considered
24
‘homogenous’, that is, the vast majority of the variance in the effect across studies can be
25

Planned behaviour and self-determination theory 12
accounted for by the statistical artifacts. Should the value fall below this criterion it is likely
1
that there is substantial variance in the effect size unattributed to methodological artifacts and
2
suggestive of variations attributable to other extraneous or ‘moderator’ variables.
3
Coding of Moderators
4
We anticipated that five key factors would moderate the proposed effects based on
5
previous research (Armitage & Conner, 2001; Hagger et al., 2002b): age of participants (old vs.
6
young), gender, publication status (published vs. unpublished), and study design (correlational
7
vs. experimental/intervention), and behaviour type (physical activity vs. other health-related
8
behaviours). The influence of moderator variables was evaluated by segregating studies on the
9
basis of the moderator variable and conducting separate meta-analyses for each moderator
10
group. The moderator was considered effective if the average corrected effect sizes calculated
11
in each moderator group were significantly different as evidenced by the CI95. Moderation was
12
further supported if the moderator resulted in a narrowing of the credibility intervals and an
13
increase in the variance accounted for by statistical artifacts.
14
We used the criteria offered by previous studies to define age moderator groups of studies
15
for older (over 18 years) and younger (under 18 years) samples (Hagger et al., 2002b; Sheeran
16
& Orbell, 1998). As only four studies reported effect sizes for exclusively female samples and
17
none tested effects in an exclusively male comparison group of participants a moderator
18
analyses by gender was not possible. While the number of unpublished studies was relatively
19
small (n = 4), it still permitted an evaluation of publication bias as a moderator variable. To test
20
the moderation of the effect sizes across studies by study design, we classified studies into
21
those that had used experimental or intervention designs to manipulate one or more of the TPB
22
or SDT variables and examined their effects on the remaining variables. Finally, although the
23
vast majority of the studies were conducted with physical activity as the focal behaviour, four
24
studies conducted analyses on other health-related behaviours, namely dieting (Hagger &
25

Planned behaviour and self-determination theory 13
Chatzisarantis, 2008a; Hagger et al., 2006), breast feeding (Wells, Thompson, & Kloeblen-
1
Tarver, 2002), and condom use (Rentzelas & Hagger, 2008), so we were therefore able to
2
compare results the effect sizes across moderator groups of physical activity and other health-
3
related behaviours.
4
Testing the Integrated Model
5
The corrected correlations derived from the meta-analysis were used as an input matrix
6
for a path analysis to test the hypotheses of the proposed model integrating the TPB and SDT
7
(Viswesvaran & Ones, 1995). As the model was based on correlations from different subsets of
8
studies there was variation in the sample sizes used to compute the corrected averaged
9
correlations for each effect. In order to reduce bias caused by the variation in sample sizes
10
across studies we opted for the most conservative strategy proposed by other researchers and
11
used the smallest sample size (Carr, Schmidt, Ford, & DeShon, 2003; Viswesvaran & Ones,
12
1995). The goodness-of-fit of the model was evaluated relative to a fully independent (totally
13
free) model. Goodness-of-fit was established using multiple criteria the Comparative Fit Index
14
(CFI) and Non-Normed Fit Index (NNFI), which should exceed .95 for a well-fitting model
15
(Hu & Bentler, 1999), and the Root Mean Square Error of Approximation (RMSEA) and the
16
Standardized Root Mean Squared Residuals (SRMSR), which should be close to .05 and .08
17
respectively for a well-fitting model (Hu & Bentler, 1999).
18
Results
19
Summary of Studies
20
Summary statistics for the studies included in the analysis are provided in Table 1. The
21
table provides all pertinent information germane to the analysis including the coding used to
22
conduct the moderator analyses.
23
Zero-Order Analysis
24

Planned behaviour and self-determination theory 14
The averaged correlation coefficients corrected for artifacts of sampling and
1
measurement error for the relationships among the TPB and SDT constructs from this sample
2
of studies are provided in Table 2. The 95% confidence intervals indicated that all of the
3
corrected correlations were significantly different from zero. The 90% credibility intervals
4
indicated that the majority of the effects exhibited discriminant validity. The only exceptions
5
were the credibility intervals for the intention-past behaviour and behaviour-past behaviour
6
correlations that included the value of 1.00. This finding was not surprising given the strong
7
prediction of the study variables and future behaviour by past behaviour. However, although it
8
is important to control for the effect of past behaviour in social cognitive models, its effect
9
merely represents the effects of unmeasured constructs rather than effects of meaningful
10
psychological variables (Ajzen, 2002; Ouellette & Wood, 1998). Overall, these findings
11
provide confirmation that the tested effects were representative of a true effect in the
12
population. Furthermore, the ‘fail safe’ sample size (NFS) values all exceeded Rosenberg’s
13
(2005) recommended critical value of 5N + 10. The only exception was the NFS for the effect
14
between perceived autonomy support and PBC, and although the CI95 value did not include the
15
value of zero, a relatively small number of null findings would overturn the significant effect.
16
Moderator Analyses
17
Moderator analyses were conducted in cases where there were more than three studies in
18
a moderator group. A summary of the average corrected effect sizes that were significantly
19
different across the moderator groups is provided in Table 3. Turning first to age as a
20
moderator, only the self-determination-PBC relationship was significantly different across
21
younger (r++ = .59) and older samples (r++ = .39). It seems that younger samples equate self-
22
determined motivation with perceived control more strongly than older samples. In terms of
23
the moderation by behaviour type, significantly stronger effects were evident in studies with
24
physical activity as the target behaviour relative to studies with other health-related behaviours
25

Planned behaviour and self-determination theory 15
as the target for the attitude-PBC (physical activity, r++ = .57; other health-related behaviours
1
r++ = .30), and PBC-behaviour (r++ = .41; .13) relationships. Conversely, the effects were
2
significantly stronger in studies with other health-related behaviours as the target behaviour
3
relative to studies with physical activity as the target for the attitude-past behaviour (physical
4
activity, r++ = .39; other health-related behaviours r++ = .70) and intention-past behaviour (r++ =
5
.52; .89) relationships. The self-determined motivation-intention relationship (correlational r++
6
= .53; experimental/intervention r++ = .32) was significantly stronger in studies adopting a
7
correlational design compared to those adopting experimental or intervention designs. It seems
8
that correlational designs tend to result in an overestimation of this effect size, although,
9
importantly, adopting an experimental or intervention design does not attenuate the proposed
10
effect to a trivial value. Finally, none of the effects were significantly moderated by publication
11
status.
12
Path Analysis
13
The meta-analytically derived corrected correlation matrix was used test the pattern of
14
relationships as stipulated by the proposed motivational sequence. The corrected averaged
15
correlations were used as an input matrix for a path analytic model that stipulated a priori the
16
proposed pattern of relationships in the sequence. The model was estimated using the EQS
17
structural equation modelling computer software using a maximum likelihood method (Bentler,
18
2004). The path model exhibited acceptable goodness-of-fit with the data according to the
19
multiple criteria adopted (CFI = .98; NNFI = .98; SRMSR = .03; RMSEA = .08). Beta
20
coefficients from the meta-analytic path analysis are provided in Figure 1. Overall, the model
21
accounted for 64.6% and 59.2% of the variance in intentions and behaviour respectively.
22
In terms of specific effects in the model, there was a statistically-significant effect of
23
perceived autonomy support on self-determined motivation ( = .31, p < .01). Self-determined
24
motivation predicted attitudes ( = .44, p < .01) and PBC ( = .38, p < .01) in accordance with
25

Planned behaviour and self-determination theory 16
the hypothesised integrated model. There was also a significant, positive relationship between
1
self-determined motivation and subjective norms ( = .14, p < .01) which was incongruent with
2
hypotheses. Attitudes ( = .37, p < .01), subjective norms ( = .06, p < .01), and PBC ( = .23,
3
p < .01) significantly predicted intentions and intentions significantly predicted behaviour ( =
4
.29, p < .01) in accordance with the TPB. There was a significant direct effect for self-
5
determined motivation on intention ( = .10, p < .01), but the size of the effect was small
6
relative to the indirect effect of self-determined motivation on intention ( = .27, p < .01).
7
Together, these effects yielded a significant total effect of self-determined motivation on
8
intention ( = .37, p < .01). There were also a significant indirect effect of self-determined
9
motivation on behaviour mediated by the motivational sequence ( = .10, p < .01), but no
10
significant direct effect3. Importantly, these were the attenuated effects after controlling for
11
past behaviour. Therefore, although past behaviour significantly predicted all of the constructs
12
in the model, it did not reduce the effects to trivial values and suggests that the motivation
13
variables in the integrated model had motivational and behavioural relevance4.
14
Discussion
15
The aim of the present study was to provide a cumulative synthesis of findings from
16
studies that have integrated two leading theories of motivation in health contexts: the theory of
17
planned behaviour (TPB) and self-determination theory (SDT). A comprehensive literature
18
search of quantitative tests of at least one effect between constructs from the TPB and SDT
19
yielded 34 studies with 43 independent tests of effects. Methods of meta-analysis proposed by
20
Hunter and Schmidt (1994) were used to produce averaged correlation coefficients among
21
constructs from the studies corrected for the methodological artifacts of sampling and
22
measurement error. All correlations were significant and only two did not exhibit discriminant
23
validity. We tested the influence of five possible moderators of the effects: age, gender,
24
publication status, study design, and behaviour type. Age significantly moderated the self-
25

Planned behaviour and self-determination theory 17
determined motivation-PBC relation, behaviour type significantly moderated the attitude-PBC,
1
attitude-past behaviour, intention-past behaviour, and PBC-behaviour relationships, and study
2
design moderated the self-determined motivation-intention relationship.
3
The hypothesized pattern of relations or ‘motivational sequence’ among the TPB and
4
SDT constructs were tested using path analysis with the meta-analytically derived correlation
5
coefficients as an input matrix. As expected, perceived autonomy support was a significant
6
predictor of self-determined motivation. Self-determined motivation significantly predicted
7
intentions to engage in health-related behaviour mediated by attitudes and perceived
8
behavioural control (PBC). Contrary to hypotheses, subjective norms was also positively
9
related to self-determined motivation which led to a significant albeit small mediated path of
10
self-determined motivation on intention through this variable. Intention mediated the effect of
11
attitudes, subjective norms, and PBC on behaviour. There was a significant indirect effect of
12
self-determined motivation on behaviour with no direct effect. Finally, the proposed pattern of
13
effects was supported when controlling for past behaviour.
14
Support for Theoretical Integration
15
The present synthesis provides support for integrated motivational approaches adopting
16
the TPB and SDT to explain health-related behaviour. The pattern of effects in the path
17
analysis support those found in individual tests of the relations among constructs of the two
18
theories and the significant mediation effects that provide insight into the possible processes
19
involved. A key effect is the prediction self-determined motivation by perceived autonomy
20
support. Perceived autonomy support serves as a proxy of social events that support self-
21
determined motivation (Deci, Eghrari, Patrick, & Leone, 1994; Reeve et al., 1999). The
22
perception of these autonomy-supportive behaviours are, not surprisingly, influential in
23
fostering self-determined motivation in health-related behaviours (Hagger et al., 2007). This is
24
important as it suggests that an autonomy-supportive environment is likely an effective means
25
to promote self-determined motivation in health contexts (Chatzisarantis & Hagger, in press).
26

Planned behaviour and self-determination theory 18
Promoting self-determined motivation is clearly an important endeavour as the evidence
1
in the present integrated model shows that self-determined motivation is linked to the proximal
2
antecedents of intentional behaviour. Results from the path analysis indicated that self-
3
determined motivation has a significant effect on the proximal predictors of intention,
4
particularly attitudes and PBC, and has a significant indirect effect on intentions via the
5
mediation of these predictors. Theoretically, this lends support to SDT in that self-determined
6
motives towards health behaviour is related to individuals reporting attitudes and perceptions
7
of control that are consistent with those self-determined motives. Previous theorists have
8
supported such relations. For example, Vallerand (2007) explicitly suggests that self-
9
determined motivation will predict cognitive beliefs regarding the target behaviour.
10
Although we included both direct and indirect effects of self-determined motivation on
11
intentions and health-related behaviour there were no significant direct effects found. This was
12
as expected and congruent with previous research (e.g., Hagger et al., 2003; Pihu, Hein, Koka,
13
& Hagger, in press; Shen et al., 2007). Previous studies have reported direct effects of self-
14
determined motivation on intentions and behaviour (e.g., Chatzisarantis et al., 2002; Hagger et
15
al., 2005; Hagger et al., in press) and it has been suggested that such effects indicate the more
16
impulsive, less deliberative processes by which self-determined motives predict intention
17
formation and enactment (Hagger et al., 2006). However, present results suggest that such
18
effects are comparatively unsubstantial compared with indirect effects. This indicates that the
19
process by which self-determined motivation affects intention formation and behavioural
20
engagement is one that is reflective rather than impulsive (Strack & Deutsch, 2004).
21
One finding that was contrary to hypotheses was the significant and positive effect of
22
self-determined motivation on subjective norms, and effect that was hypothesised to be zero or
23
even negative. This is because subjective norms are defined as social pressures to engage in
24
future behaviour and therefore reflect more controlling, externally-referenced beliefs about
25
engaging in future health behaviour. However, subjective norms may, on the surface, reflect
26

Planned behaviour and self-determination theory 19
beliefs about the controlling nature of others, but these ‘others’ are generally specified as
1
‘significant others’ in measures of the construct. Research in SDT has suggested that people
2
will tend to conform to the wishes of significant others because although they may appear
3
outwardly controlling, if the referent is a significant other, then the person is likely to have
4
internalised their demands as supportive of their self-determined motivation (Iyengar &
5
Lepper, 1999). In other words, the person perceives the significant other as acting in their best
6
interests and supporting their autonomy. As a consequence, the variable may reflect both
7
controlling and internalised aspects of social beliefs on future health behaviour engagement.
8
This may account for the inconsistent findings for this relationship in the literature (Hagger et
9
al., 2005; Hagger et al., 2003). Future studies should make the distinction between perceived
10
pressure from different referents at the beliefs level as recommended by Ajzen (1985) and
11
relate them to self-determined motivation. This would test the hypothesis that self-determined
12
motivation would be positively related to beliefs about salient referents’ social influence while
13
a negative relationship would be expected for beliefs about non-salient referents’ influence.
14
Importance of Moderators
15
In the present study, this was relevant as none of the effect sizes could be considered
16
homogenous according to the Hunter and Schmidt (1994) 75% rule indicating the presence of
17
moderators. The moderation analyses produced some interesting contrasts. For example, the
18
relationship between self-determined motivation and PBC was significantly stronger in studies
19
with younger samples. This may be due to the fact that younger people have a less-
20
differentiated view of self-determined motivation and may equate it more with competence
21
than older samples, as evidenced by developmental tests of self-determined motivation (Otis,
22
Grouzet, & Pelletier, 2005). Behaviour type also moderated a number of effects most notably
23
the attitude-PBC and PBC-behaviour relationships. This may be because the behaviours
24
included in the other health-related behaviours sample (e.g., dieting and condom use) are
25

Planned behaviour and self-determination theory 20
behaviours for which there is a strong social component compared with physical activity which
1
has stronger attitudinal and control components (Hagger et al., 2006; Sheeran & Orbell, 1998).
2
Limitations and Suggestions for Future Research
3
A limitation of studies integrating these theories is its heavy bias towards cross-sectional
4
and prospective designs and a disproportionate focus on physical activity as the target
5
behaviour. Current studies using this integrative approach have adopted valid and reliable
6
measurement instruments (e.g., Hagger et al., 2007; Markland & Tobin, 2004) with large,
7
representative samples and appropriate prospective designs to avoid confounding artifacts like
8
common method variance (e.g., Hagger et al., 2003). However, the majority of the studies
9
included in the analysis were correlational in design, which prevents the inference of causality
10
(Weinstein, 2007). Furthermore, although we have articulated a theory to account for the
11
proposed sequence of effects in the integration of the SDT and TPB, there are likely to be
12
alternative models that exhibit good fit with the data, although such models would need
13
theoretical justification to illustrate their plausibility. On the issue of causality, the moderator
14
analyses using study design as a moderator provided some preliminary evidence that findings
15
from the few experimental or intervention designs included in the present study were largely
16
consistent with those of the correlational studies. Not only does this suggest that the method
17
adopted, on the whole, does not result in a drastic attenuation of the effect sizes in the proposed
18
integrated theory, but it also provides some preliminary evidence of the causal nature of the
19
hypothesised relationships in the theory. However, there is a need for further research to bolster
20
support for the integrative approach advocated here by adopting experimental or randomised
21
controlled intervention designs to manipulate SDT constructs and examining their influence on
22
TPB variables (e.g., Chatzisarantis & Hagger, in press). Such designs will permit researchers
23
integrating these theories to generate evidence that either corroborates or refutes the proposed
24
direction of causality among the variables.
25

Planned behaviour and self-determination theory 21
A further limitation is that the evaluation of moderators was unable to resolve the
1
heterogeneity in the meta-analysed effect sizes. Therefore, the path analytic model was
2
estimated using effect sizes that may vary due to the influence of moderator variables. The
3
problem is exacerbated by a general difficulty in modelling such interactions in analyses based
4
on structural equations as this is known to produce biased estimates (Viswesvaran & Ones,
5
1995). These difficulties also apply to individual studies and there is a need for future research
6
to test moderation effects of the effects in mediation models (Harris & Hagger, 2007).
7
We have indicated in the current analysis of the probable presence of moderator variables
8
and tested a limited set based on previous research to ensure that our analysis was informative
9
as possible and to identify directions for future research. However, the potential for moderators
10
to affect the true nature of the effects proposed by path analysis is a real limitation and has
11
been acknowledged in previous studies adopting the same approach (Carr et al., 2003; Colquitt
12
& LePine, 2000; Schepers & Wetzels, 2007). Although we have provided some evidence based
13
on the confidence intervals of the meta-analysed correlations that the heterogeneity has little
14
effect on the overall pattern of results, this does not resolve the problem that some paths in the
15
model may be moderated by extraneous variables. More research studies integrating these
16
theories may permit future path analytic meta-analyses to be conducted separately across sets
17
of studies by moderator subgroups using a multi-group procedure (Edwards & Lambert, 2007;
18
Hom, Caranikas-Walker, Prussia, & Griffeth, 1992), although given the relative dearth of
19
studies conducting analyses on moderator groups, this will be unlikely in the very near future.
20
Finally, we were only able to locate three studies integrating the TPB and SDT on health-
21
related behaviours other than health-related physical activity. Although type of health-related
22
behaviour did not serve to moderate the proposed relations among the effects in the present
23
analysis, future research needs to examine these effects in a wider range of health behaviour
24
domains to further corroborate the consistency of these effects.
25

Planned behaviour and self-determination theory 22
References
* Denotes studies included in meta-analysis
Ajzen, I. (1985). From intentions to actions: A theory of planned behavior. In J. Kuhl & J.
Beckmann (Eds.), Action-control: From cognition to behavior (pp. 11-39). Heidelberg:
Springer.
Ajzen, I. (2002). Residual effects of past on later behavior: Habituation and reasoned action
perspectives. Personality and Social Psychology Review, 6, 107-122.
Ajzen, I. (2003). Constructing a TPB questionnaire: Conceptual and methodological
considerations. Retrieved April 14, 2003, from University of Massachusetts,
Department of Psychology Web site: http://www-unix.oit.umass.edu/~aizen: University
of Massachusetts.
Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behavior. New
Jersey: Prentice Hall.
*Alexandris, K., Kouthouris, C., & Girgolas, G. (2007). Investigating the relationships among
motivation, negotiation, and alpine skiing participation. Journal of Leisure Research,
39, 648-667.
Andersen, S. M., Chen, S., & Carter, C. (2000). Fundamental human needs: Making social
cognition relevant. Psychological Inquiry, 4, 269-275.
Armitage, C. J., & Conner, M. (2001). Efficacy of the theory of planned behaviour: A meta-
analytic review. British Journal of Social Psychology, 40, 471-499.
Baard, P. P., Deci, E. L., & Ryan, R. M. (2004). Intrinsic need satisfaction: A motivational
basis of performance and well-being in two work settings. Journal of Applied Social
Psychology, 34, 2045-2068.
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social
psychological research: Conceptual, strategic and statistical considerations. Journal of
Personality and Social Psychology, 51, 1173-1182.
Bentler, P. M. (2004). EQS structural equations modeling software (Version 6.1) [Computer
Software]. Encino, CA: Multivariate Software.
Biddle, S. J. H., Cury, F., Goudas, M., Sarrazin, P., Famose, J. P., & Durand, M. (1995).
Development of scales to measure perceived physical education class climate: A cross-
national project. British Journal of Educational Psychology, 65, 341-358.
*Biddle, S. J. H., Soos, I., & Chatzisarantis, N. (1999). Predicting physical activity intentions
using goal perspectives and self-determination theory approaches. European
Psychologist, 4, 83-89.

Planned behaviour and self-determination theory 23
*Brickell, T., Chatzisarantis, N. L. D., & Pretty, G. M. (2006). Autonomy and control:
Augmenting the validity of the Theory of Planned Behaviour in predicting exercise
Journal of Health Psychology, 11, 51-63.
Briere, N., Vallerand, R. J., Blais, N., & Pelletier, L. (1995). Development and validation of a
measure of intrinsic, extrinsic, and amotivation in sports: The Sport Motivation Scale
(SMS). International Journal of Sport Psychology, 26, 465-489.
Carr, J. Z., Schmidt, A. M., Ford, J. K., & DeShon, R. P. (2003). Climate perceptions matter: A
meta-analytic path analysis relating molar climate, cognitive and affective states, and
individual level work outcomes. Journal of Applied Psychology, 88, 605-619.
*Chatzisarantis, N. L. D., & Hagger, M. S. (in press). Effects of an intervention based on Self-
Determination Theory on self-reported leisure-time physical activity participation.
Psychology and Health.
*Chatzisarantis, N. L. D., Hagger, M. S., Biddle, S. J. H., & Karageorghis, C. (2002). The
cognitive processes by which perceived locus of causality predicts participation in
physical activity. Journal of Health Psychology, 7, 685-699.
Chatzisarantis, N. L. D., Hagger, M. S., Biddle, S. J. H., Smith, B., & Wang, J. C. K. (2003). A
meta-analysis of perceived locus of causality in exercise, sport, and physical education
contexts. Journal of Sport and Exercise Psychology, 25, 284-306.
*Chatzisarantis, N. L. D., Hagger, M. S., & Smith, B. (2007). Influences of perceived
autonomy support on physical activity within the theory of planned behavior. European
Journal of Social Psychology, 37, 934-954.
*Chatzisarantis, N. L. D., Hagger, M. S., Smith, B., & Sage, L. D. (2006). The influences of
intrinsic motivation on execution of social behaviour within the theory of planned
behaviour. European Journal of Social Psychology, 36, 229-237.
Cheung, G. W., & Rensfold, R. B. (2002). Evaluating goodness-of-fit indexes for testing
measurement invariance. Structural Equation Modeling, 9, 233-255.
Colquitt, J. A., & LePine, J. A. (2000). Toward an integrative theory of training motivation: A
meta-analytic path analysis of 20 years of research. Journal of Applied Psychology, 85,
678-707.
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.
Conner, M., & Norman, P. (2005). Predicting health behaviour: Research and practice with
social cognition models. Buckingham, UK: Open University Press.

Planned behaviour and self-determination theory 24
Deci, E. L., Eghrari, H., Patrick, B. C., & Leone, D. R. (1994). Facilitating internalization: The
self-determination theory perspective. Journal of Personality, 62, 119-142.
Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and self-determination in human
behavior. New York: Plenum Press.
Deci, E. L., & Ryan, R. M. (2000). The "What" and "Why" of goal pursuits: Human needs and
the self-determination of behavior. Psychological Inquiry, 11, 227-268.
*Edmunds, J. K., Ntoumanis, N., & Duda, J. L. (2007). Adherence and well-being in
overweight and obese patients referred to an exercise prescription scheme: A self-
determination theory perspective. Psychology of Sport and Exercise, 8, 722-740.
*Edmunds, J. K., Ntoumanis, N., & Duda, J. L. D. (in press). Testing a self-determination
theory based teaching style in the exercise domain. European Journal of Social
Psychology.
Edwards, J. R., & Lambert, L. S. (2007). Methods for integrating moderation and mediation: A
general analytical framework using moderated path analysis. Psychological Methods,
12(1), 1-22.
Field, A. P. (2001). Meta-analysis of correlation coefficients: A Monte Carlo comparison of
fixed- and random- effects methods. Psychological Methods, 6, 161-180.
Godin, G., & Shephard, R. J. (1985). A simple method to assess exercise behavior in the
community. Canadian Journal of Applied Sport Science, 10, 141-146.
*Goudas, M., Biddle, S. J. H., & Fox, K. R. (1994). Perceived locus of causality, goal
orientations, and perceived competence in school physical education classes. British
Journal of Educational Psychology, 64, 453-563.
Goudas, M., Dermitzaki, I., & Bagiatis, K. (2001). Motivation in physical education is
correlated with participation in sport after school. Psychological Reports, 88, 491-496.
Hagger, M. S. (2006). Meta-analysis in sport and exercise research: Review, recent
developments, and recommendations. European Journal of Sport Science, 6, 103-115.
*Hagger, M. S., & Chatzisarantis, N. L. D. (2008a). Influence of impulsivity in the theory of
planned behaviour and self-determination theory for dieting behaviours. Unpublished
manuscript, University of Nottingham, Nottingham, UK.
Hagger, M. S., & Chatzisarantis, N. L. D. (2008b). Self-determination theory and the
psychology of exercise. International Review of Sport and Exercise Psychology, 1, 79-
103.
*Hagger, M. S., Chatzisarantis, N. L. D., Barkoukis, V., Wang, C. K. J., & Baranowski, J.
(2005). Perceived autonomy support in physical education and leisure-time physical

Planned behaviour and self-determination theory 25
activity: A cross-cultural evaluation of the trans-contextual model. Journal of
Educational Psychology, 97, 376-390.
*Hagger, M. S., Chatzisarantis, N. L. D., & Biddle, S. J. H. (2002a). The influence of
autonomous and controlling motives on physical activity intentions within the Theory
of Planned Behaviour. British Journal of Health Psychology, 7, 283-297.
Hagger, M. S., Chatzisarantis, N. L. D., & Biddle, S. J. H. (2002b). 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 and Exercise
Psychology, 24, 3-32.
*Hagger, M. S., Chatzisarantis, N. L. D., Culverhouse, T., & Biddle, S. J. H. (2003). The
processes by which perceived autonomy support in physical education promotes
leisure-time physical activity intentions and behavior: A trans-contextual model.
Journal of Educational Psychology, 95, 784–795.
*Hagger, M. S., Chatzisarantis, N. L. D., & Harris, J. (2006). From psychological need
satisfaction to intentional behavior: Testing a motivational sequence in two behavioral
contexts. Personality and Social Psychology Bulletin, 32, 131-138.
Hagger, M. S., Chatzisarantis, N. L. D., Hein, V., Pihu, M., Soós, I., & Karsai, I. (2007). The
perceived autonomy support scale for exercise settings (PASSES): Development,
validity, and cross-cultural invariance in young people. Psychology of Sport and
Exercise, 8, 632-653.
*Hagger, M. S., Chatzisarantis, N. L. D., Hein, V., Pihu, M., Soós, I., Karsai, I., Lintunen, T.,
& Leemans, S. (in press). Teacher, peer, and parent autonomy support in physical
education and leisure-time physical activity: A trans-contextual model of motivation in
four cultures. Psychology and Health.
Harris, J., & Hagger, M. S. (2007). Do basic psychological needs moderate relationships within
the theory of planned behavior? Journal of Applied Biobehavioral Research, 12, 43-64.
Hom, P. W., Caranikas-Walker, F., Prussia, G. E., & Griffeth, R. W. (1992). A meta-analytical
structural equations analysis of a model of employee turnover. Journal of Applied
Psychology, 77, 890-909.
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis:
Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1-55.
Hunter, J. E., & Schmidt, F. (1994). Methods of meta-analysis: Correcting error and bias in
research findings (2nd ed.). Newbury Park, CA: Sage.

Planned behaviour and self-determination theory 26
Hunter, J. E., & Schmidt, F. L. (2000). Fixed effects vs. random effects meta-analysis models:
Implications for cumulative research knowledge in psychology. International Journal
of Selection and Assessment, 8, 275-292.
Iyengar, S. S., & Lepper, M. R. (1999). Rethinking the value of choice: A cultural perspective
on intrinsic motivation. Journal of Personality and Social Psychology, 76, 349-366.
Johnston, M. (2005). Means and meanings in health psychology. European Health
Psychologist, 1, 6-8.
Lin, H. F. (2007). Effects of extrinsic and intrinsic motivation on employee knowledge sharing
intentions. Journal of Information Science, 33, 135-149.
Markland, D., & Tobin, V. (2004). A modification to the Behavioural Regulation in Exercise
Questionnaire to include an assessment of amotivation. Journal of Sport and Exercise
Psychology, 26, 191-196.
*Martin Ginis, K. A., Jung, M. E., Brawley, L. R., Latimer, A. E., Hicks, A. L., Shields, C. A.,
& McCartney, N. (2006). The effects of physical activity enjoyment on sedentary older
adults' physical activity attitudes and intentions Journal of Applied Biobehavioral
Research, 11, 29-43.
McAuley, E., Duncan, T., & Tammen, V. V. (1989). Psychometric properties of the intrinsic
motivation inventory in a competitive sport setting: A confirmatory factor analysis.
Research Quarterly for Exercise and Sport.
*McLachlan, S., & Hagger, M. S. (2008). Implicit and explicit reasons for participating in
physical activity. Unpublished manuscript, University of Nottingham, Nottingham, UK.
Mullan, E., Markland, D., & Ingledew, D. K. (1997). A graded conceptualisation of self-
determination in the regulation of exercise behaviour: Development of a measure using
confirmatory factor analysis. Personality and Individual Differences, 23, 745-752.
*Ntoumanis, N. (2001). A self-determination approach to the understanding of motivation in
physical education. British Journal of Educational Psychology, 71, 225-242.
*Ntoumanis, N. (2005). A prospective study of participation in optional school physical
education based on self-determination theory. Journal of Educational Psychology, 97,
444-453.
Orbell, S. (2004). Intention-behaviour relations: A self-regulation perspective. In G. Haddock
& G. R. O. Maio (Eds.), Contemporary perspectives on the psychology of attitudes (pp.
145-168). London: Psychology Press.

Planned behaviour and self-determination theory 27
Otis, N., Grouzet, F. M. E., & Pelletier, L. G. (2005). Latent motivational change in an
academic setting: A 3-year longitudinal study. Journal of Educational Psychology, 97,
170-183.
Ouellette, J. A., & Wood, W. (1998). Habit and intention in everyday life: The multiple
processes by which past behavior predicts future behavior. Psychological Bulletin, 124,
54-74.
*Palmeira, A. L., Teixeira, P. J., Branco, T. L., Martins, S. S., Minderico, C. S., Barata, J. T., et
al. (2007). Predicting short-term weight loss using four leading health behavior change
theories. International Journal of Behavioral Nutrition and Physical Activity, 4, 14.
*Papacharisis, V., Simou, K., & Goudas, M. (2003). The relationship between intrinsic
motivation and intention towards exercise. Journal of Human Movement Studies, 45,
377-386.
Phillips, P., Abraham, C., & Bond, R. (2003). Personality, cognition, and university students'
examination performance. European Journal of Personality, 17, 435-448.
*Pihu, M., Hein, V., Koka, A., & Hagger, M. S. (in press). How students’ perceptions of
teacher’s autonomy-supportive behaviours affect physical activity behaviour: An
application of trans-contextual model. European Journal of Sport Science.
Reeve, J., Bolt, E., & Cai, Y. (1999). Autonomy-supportive teachers: How they teach and
motivate students. Journal of Educational Psychology, 91, 537-548.
*Rentzelas, P., & Hagger, M. S. (2008). Influence of independence and interdependence as
moderators of relationships in the theory of planned behaviour. Unpublished
manuscript, University of Nottingham, Nottingham, UK.
Rosenberg, M. S. (2005). The file-drawer problem revisited: A general weighted method for
calculating fail-safe numbers in meta-analysis. Evolution 59, 464-468.
Rosenthal, M. C. (1994). The fugitive literature. In H. Cooper & L. V. Hedges (Eds.), The
Handbook of Research Synthesis (pp. 85-94). New York: Russell Sage Foundation.
Ryan, R. M., & Connell, J. P. (1989). Perceived locus of causality and internalization:
Examining reasons for acting in two domains. Journal of Personality and Social
Psychology, 57, 749-761.
*Sarrazin, P., Vallerand, R. J., Guillet, E., Pelletier, L. G., & Cury, F. (2002). Motivation and
dropout in female handballers: A 21-month prospective study. European Journal of
Social Psychology, 32, 395-418.

Planned behaviour and self-determination theory 28
Schepers, J., & Wetzels, M. (2007). A meta-analysis of the technology acceptance model:
Investigating subjective norm and moderation effects. Information & Management,
44(1), 90-103.
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.
*Shen, B., McCaughtry, N., & Martin, J. (2007). The influence of self-determination in
physical education on leisure-time physical activity behavior Research Quarterly for
Exercise and Sport, 78, 328-338.
Sobel, M. E. (1982). Asymptotic confidence intervals for indirect effects in structural equation
models. In S. Leinhardt (Ed.), Sociological Methodology (pp. 290-312). San Francisco,
CA: Jossey-Bass.
*Standage, M., Duda, J. L., & Ntoumanis, N. (2003). A model of contextual motivation in
physical education: Using constructs from self-determination and achievement goal
theories to predict physical activity intentions. Journal of Educational Psychology, 95,
97-110.
Stauffer, J. M. (1996). A Graphical User Interface psychometric meta-analysis program for
DOS Educational and Psychological Measurement, 56, 675-677.
Strack, F., & Deutsch, R. (2004). Reflective and impulsive determinants of social behavior.
Personality and Social Psychology Review, 8, 220-247.
*Thøgersen-Ntoumani, E. C., & Ntoumanis, N. (2007). The role of self-determined motivation
in the understanding of exercise-related behaviours, cognitions and physical self-
evaluations. Journal of Sports Sciences, 24, 393-404.
Vallerand, R. J. (1997). Towards a hierarchical model of intrinsic and extrinsic motivation. In
M. P. Zanna (Ed.), Advances in experimental social psychology (pp. 271-359). New
York: Academic Press.
Vallerand, R. J. (2007). A hierarchical model of intrinsic and extrinsic motivation for sport and
physical activity. In M. S. Hagger & N. L. D. Chatzisarantis (Eds.), Intrinsic Motivation
and Self-Determination in Exercise and Sport (pp. 255-279). Champaign, Il: Human
Kinetics.
*Vierling, K. K., Standage, M., & Treasure, D. C. (2007). Predicting attitudes and physical
activity in an "at-risk" minority youth sample: A test of self-determination theory.
Psychology of Sport and Exercise, 8, 795-817.

Planned behaviour and self-determination theory 29
Viswesvaran, C., & Ones, D. S. (1995). Theory testing: Combining psychometric meta-
analysis and structural equations modeling. Personnel Psychology, 48, 865-886.
*Vlachopoulos, S., Karageorghis, C., & Terry, P. C. (2000). Motivation profiles in sport: A
self-determination theory perspective. Research Quarterly for Exercise and Sport, 71,
387-397.
*Wallhead, T. L., & Hagger, M. S. (2008). The trans-contextual model in Caucasian and
American Indian school children. Unpublished manuscript, University of Wyoming,
Laramie, WY.
Washburn, R. A., Smith, K. W., Jette, A. M., & Janney, C. A. (1993). The Physical Activity
Scale for the Elderly (PASE): Development and evaluation. Journal of Clinical
Epidemiology, 46, 153-162.
Weinstein, N. D. (2007). Misleading tests of health behavior theories. Annals of Behavioral
Medicine, 33, 1-10.
*Wells, K. J., Thompson, N. J., & Kloeblen-Tarver, A. S. (2002). Intrinsic and extrinsic
motivation and intention to breast-feed. American Journal of Health Behavior, 26, 111-
120.
Williams, G. C., Cox, E. M., Kouides, R., & Deci, E. L. (1999). Presenting the facts about
smoking to adolescents: The effects of an autonomy supportive style. Archives of
Pediatrics and Adolescent Medicine, 153, 959-964.
Williams, G. C., & Deci, E. L. (1996). Internalization of biopsychosocial values by medical
students: A test of self-determination theory. Journal of Personality and Social
Psychology, 70, 767-779.
*Wilson, P. M., & Rodgers, W. M. (2004). The relationship between perceived autonomy
support, exercise regulations and behavioral intentions in women. Psychology of Sport
and Exercise, 5, 229-242.
*Wilson, P. M., Rodgers, W. M., Blanchard, C. M., & Gessell, J. (2003). The relationship
between psychological needs, self-determined motivation, exercise attitudes, and
physical fitness. Journal of Applied Social Psychology, 33, 2373-2392.

Planned behaviour and self-determination theory 30
Footnote
1Lists of all databases searched, keywords used in the searches, and journals and review
articles consulted are available from the first author on request.
2Ryan and Connell’s (1989) perceived locus of causality scale measures four types of
motivation, each varying in the degree of self-determination on a continuum ranging from high
to low self-determination, known as the perceived locus of causality (PLOC). Most self-
determined is intrinsic motivation, which is the prototypical form of autonomous motivation
representing behavioural engagement for no external contingency or reinforcement. There are
three forms of extrinsic motivation, which vary in the degree to which they are self-determined
or controlled: identified regulation is a highly autonomous form of motivation representing
motivation to engage in a behaviour because it services goals that are intrinsic and salient to the
self, introjected regulation is a less autonomous form of motivation reflecting behavioural
engagement due to perceived internal pressures like avoiding shame or guilt or gaining
contingent self-worth or pride and external regulation is the prototypical form of extrinsic
motivation, and therefore the least self-determined, reflecting engaging in behaviours due to
external reinforcement such as obtaining a reward or avoiding punishment. In the present
study, we used either unidimensional measures of self-determined motivation, the weighted
self-determination index based on the PLOC or individual self-determined scales from the
perceived locus of causality index to as equivalent measures of self-determined motivation.
There were insufficient studies to explore individual tests of the individual perceived locus of
causality subscales with TPB variables. In cases where studies did not report the reweighted
self-determination index but reported effect sizes between the two most self-determined forms
of motivation from the PLOC, namely, intrinsic motivation and identified regulation, and the
TPB variables, we used the arithmetic average value of the two effect sizes in our analysis.
3In all analyses testing for significant indirect effects the following criteria proposed by
Baron and Kenny (1986) were met: (1) significant correlations between the dependent variable

Planned behaviour and self-determination theory 31
and the independent (predictor) variable(s); (2) significant correlations between the mediator
and the independent variable(s); (3) a significant unique effect of the mediator on the
dependent variable when it is included alongside the independent variable(s) in a multivariate
test of these relationships; and (4) the significant effect of independent variable on the
dependent is attenuated or extinguished when the mediator is included as an independent
predictor of the dependent variable. The tests of indirect effects are equivalent to the Sobel
(1982) test.
4The path analysis was conducted on a matrix of meta-analysed correlations that included
considerable variance across studies unaccounted for by the methodological artifacts for which
the effect sizes were corrected. This suggested the presence of moderator variables.
Viswesvaren and Ones (1995) cite this as one drawback of using meta-analysed effect sizes
using path analyses. One solution they offer is to test the relative contribution of the
heterogeneity makes to the pattern of relations in the model and the overall goodness-of-fit of
the model. They proposed that the fit of the model estimated using the mean corrected
correlation coefficients be compared with the same model estimated using the upper bound and
lower bound confidence intervals (CI95) of the corrected correlations using multi-group path
analyses. We therefore estimated both these models in the present meta-analysis. We first
estimated a baseline multi-group model in which the tenability of the pattern of relationships
was tested between the model based on the mean corrected correlation matrix and models
based on the upper and lower bound CI95 values. We subsequently estimated a constrained
multi-group model in which equality constraint equations were specified to test the invariance
of the specified paths among the model variables across models based on the upper and lower
bound confidence intervals of the corrected correlations and the model based on the mean
corrected correlations. Results indicated that the effect of the heterogeneity as represented by
the confidence intervals of the effect sizes did not produce substantial deviations in the CFI and
NNFI fit indexes according to the criteria specified by Cheung and Rensvold (2002) of a

Planned behaviour and self-determination theory 32
difference in the indexes of less than .01. This provides some evidence that heterogeneity in the
relations does not contribute substantially to variations in the pattern of effects specified in the
model based on the mean correlations.

Planned behaviour and self-determination theory 33
Table 1 Continues
Table 1.
Characteristics of Studies Included in Meta-Analysis of Theory of Planned Behaviour and Self-Determination Theory
Study
Health
Behaviour
Sample(s)1
Mean Age of
Sample (SD)
TPB Constructs
Measured
SDT Constructs Measured2
Study Design
Behaviour
Measure
Past
Behaviour
Measure
Alexandris,
Kouthouris, & Grouios
(2007) [A]
Physical
activity
[A]
220 (92 females, 128 males)
adults
Range: 18-40
years [B]
Intention
SMS
Intrinsic (know),
Intrinsic (accomplishment),
Intrinsic (stimulation),
Identified,
Introjection,
External
Correlational, single
wave [A]
—
—
Biddle, Soos, &
Chatzisarantis (1999)
[A]
Physical
activity
[A]
723 high school pupils
Range: 12-16
years [A]
Intention
PLOC
Intrinsic,
Identified,
Introjection,
External
Correlational, single
wave [A]
—
—
Brickell,
Chatzisarantis, &
Pretty (2006) [A]
Physical
activity
[A]
163 (99 females, 63 males)
University students and staff
23.15 (6.05)
Range: 18-44
[B]
All
Autonomous and controlling intentions
Correlational,
prospective – 2 weeks
[A]
Self-report –
LTEQ
—
Chatzisarantis &
Hagger (in press) [A]
Physical
activity
[A]
215 (109 females, 106 males)
high school pupils
14.84 (0.48)
All
PLOC
Intrinsic,
Identified,
Introjection,
External;
Perceived autonomy support (LCQ)
Intervention based on
SDT – baseline and 5-
week follow-up
measures [B]
Self-report –
LTEQ
Self-report –
1 item
Chatzisarantis,
Hagger, Biddle, &
Karageorghis (2002)
[A]
Physical
activity
[A]
168 (83 females, 85 males)
high school pupils
13.53 (0.05)
[A]
Intention,
Attitude,
PBC
PLOC
Intrinsic,
Identified,
Introjection,
External
Correlational,
prospective – 2 weeks
[A]
Self-report –
LTEQ
Self-report –
1 item
Chatzisarantis,
Hagger, & Smith
(2007) (study 1) [A]
Physical
activity
[A]
177 (107 females, 69 males)
high school pupils and
University students
School pupils:
13.95 (0.61);
University
students 18.98
(2.63) [U]
All
Perceived autonomy support (HCCQ)
Correlational,
prospective – 5 weeks
[A]
Self-report –
LTEQ
Self-report –
1 item
Chatzisarantis,
Hagger, & Smith
(2007) (study 2) [A]
Physical
activity
[A]
165 (79 females, 86 males)
high school pupils
14.56 (0.77)
[A]
All
Perceived autonomy support (HCCQ)
Correlational,
prospective – 5 weeks
[A]
Self-report –
LTEQ
Self-report –
1 item
Chatzisarantis,
Hagger, & Smith
(2007) (study 3) [A]
Physical
activity
[A]
79 (39 females, 40 males) high
school pupils
14.53 (0.70)
[A]
Intention,
Attitudes
(dependent
measures)
Perceived autonomy support (HCCQ) (dependent
measure)
Experimental, 1-way
factorial design
(autonomy-support
and control) [B]
—
—

Planned behaviour and self-determination theory 34
Table 1 Continues
Study
Health
Behaviour
Sample(s)1
Mean Age of
Sample (SD)
TPB Constructs
Measured
SDT Constructs Measured
Study Design
Behaviour
Measure
Past
Behaviour
Measure
Chatzisarantis,
Hagger, Smith, &
Sage (2006) [A]
Physical
activity
[A]
460 (254 females, 206 males)
high school pupils, University
students and adults
School pupils:
14.25 (1.04);
University
students 19.52
(1.44); adults
34.33 (1.14)
[U]
All
Intrinsic motivation (BREQ)
Correlational,
prospective – 5 weeks
[A]
Self-report –
LTEQ
Self-report –
1 item
Edmunds, Ntoumanis,
& Duda (2007)[A]
Physical
activity
[A]
49 (41 females, 8 males) obese
or overweight volunteers
44.98 (14.61)
[B]
Intention
PLOC (BREQ-2)
Intrinsic,
Identified,
Introjection,
External;
Perceived autonomy support (adapted HCCQ)
Behavioural
intervention design in
obese and overweight
people with 1- and 3-
month follow-ups [B]
Self-report –
LTEQ
—
Edmunds, Ntoumanis,
& Duda (in press) [A]
Physical
activity
[A]
56 female University students
and staff (31 in control group,
25 in SDT group)
Control group:
21.26 (3.80)
Range: 18-53;
SDT group:
21.36 (6.71)
Range: 18-38
[B]
Intention
PLOC (BREQ-2)
Intrinsic,
Identified,
Introjection,
External
SDT intervention with
5- and 9-week follow-
ups [B]
Exercise
class
attendance
—
Goudas, Dermitzaki,
& Bagiatis (2001) [A]
Physical
activity
[A]
247 (103 females, 144 males)
high school pupils
15.30 [A]
Attitude
(outcome
expectancies/
outcome
evaluation)
IMI
Enjoyment/interest
Effort/importance
Competence
Tension/Pressure
Correlational, single
wave [A]
Self-report
sports
participation
—
Hagger &
Chatzisarantis (2008a)
[B]
Dieting
[B]
153 (115 females, 53 males)
University students and staff
23.60 (10.21)
[B]
All
PLOC
Intrinsic,
Identified,
Introjection,
External
Correlational,
prospective – 4 weeks
[A]
Self-report –
2 items
Self-report –
2 items
Hagger,
Chatzisarantis, &
Biddle, (2002a) [A]
Physical
activity
[A]
1088 (537 females, 551 males)
high school pupils
Range 12-14
years [A]
All
PLOC
Intrinsic,
Identified,
Introjection,
External
Correlational, single
wave [A]
—
—
Hagger,
Chatzisarantis,
Barkoukis, Wang, &
Baranowski (2005)
[A]
Physical
activity
[A]
551 (298 females, 253 males);
British sample 222 (118
females, 104 males); Greek
sample 93 (57 females, 36
males); Polish sample 103 (56
females, 47 males);
Singaporean sample 133 (67
females, 66 males) high school
pupils
British: 14.68
(1.47); Greek:
13.99 (0.80);
Polish: 16.28
(1.12);
Singaporean:
13.32 (0.47)
[A]
All
PLOC
Intrinsic,
Identified,
Introjection,
External;
Perceived autonomy support (PASSES)
Correlational,
prospective – 5 weeks
[A]
Self-report –
LTEQ
Self-report –
1 item

Planned behaviour and self-determination theory 35
Table 1 Continues
Study
Health
Behaviour
Sample(s)1
Mean Age of
Sample (SD)
TPB Constructs
Measured
SDT Constructs Measured
Study Design
Behaviour
Measure
Past
Behaviour
Measure
Hagger,
Chatzisarantis,
Culverhouse, & Biddle
(2003) [A]
Physical
activity
[A]
295 (163 females, 132 males)
high school pupils
14.50 (1.35)
Range 13-16
[A]
All
PLOC
Intrinsic,
Identified,
Introjection,
External;
Perceived autonomy support (PASSES)
Correlational,
prospective – 5 weeks
[A]
Self-report –
LTEQ
Self-report –
1 item
Hagger,
Chatzisarantis, &
Harris (2006) (Sample
1) [A]
Physical
activity
[A]
261 (166 females, 95 males)
University students
24.93 (9.69)
[B]
All
PLOC
Intrinsic,
Identified,
Introjection,
External
Correlational,
prospective – 2 weeks
[A]
Self-report –
2-items
—
Hagger,
Chatzisarantis, &
Harris (2006) (Sample
2) [A]
Dieting
[B]
250 (141 females, 109 males)
University students
24.64 (6.39)
[B]
All
PLOC
Intrinsic,
Identified,
Introjection,
External
Correlational,
prospective – 2 weeks
Self-report –
2 items
—
Hagger,
Chatzisarantis, Hein,
Soós, Lintunen &
Leemans (in press) [A]
Physical
activity
[A]
840 (460 females, 382 males);
British sample 210 (116
females, 94 males); Estonian
sample 268 (151 females, 117
males); Finnish sample 127
(72 females, 55 males);
Hungarian sample 235 (121
females, 114 males) high
school pupils
British: 13.19
(1.12);
Estonian:
15.04 (0.91);
Finnish: 14.30
(0.49);
Hungarian:
14.02 (0.99)
[A]
All
PLOC
Intrinsic,
Identified,
Introjection,
External;
Perceived autonomy support (PASSES)
Correlational,
prospective – 5 weeks
[A]
Self-report –
LTEQ
Self-report –
1 item
Martin Ginis, Jung,
Brawley, Latimer, &
Hicks (2006) [A]
Physical
activity
[A]
41 sedentary older adults (34
females, 7 males)
75.4 (5.40) [B]
Intention
Attitude
Enjoyment
Non-theory based
intervention using
weight training and
activities of daily
living [A]
Self-report –
PASE
—
McLachlan & Hagger
(2008) [B]
Physical
activity
[A]
185 (129 females, 56 males)
adults
30.83 (13.21)
[B]
All
PLOC
Intrinsic,
Identified,
Introjection,
External
Correlational,
prospective – 3 weeks
[A]
Self-report –
2 items
Self-report –
2 items
Ntoumanis (2001) [A]
Physical
activity
[A]
428 (218 females, 206 males,
4 non-respondent) high school
pupils
14.84 (0.52)
Range 14-16
[A]
Intention
PLOC
Intrinsic,
Identified,
Introjection,
External;
Choice climate (PECCS)
Correlational – single
wave [A]
—
—
Ntoumanis (2005) [A]
Physical
activity
[A]
302 (91 females, 211 males)
high school pupils
15.00 [A]
Intention
PLOC (SRQ)
Intrinsic,
Identified,
Introjection,
External;
Perceived autonomy support (LCQ)
Correlational – single
wave [A]
—
—

Planned behaviour and self-determination theory 36
Table 1 Continues
Study
Health
Behaviour
Sample(s)1
Mean Age of
Sample (SD)
TPB Constructs
Measured
SDT Constructs Measured
Study Design
Behaviour
Measure
Past
Behaviour
Measure
Palmeira, Teixeira,
Branco, Martins,
Minderico, Barata,
Serpa, & Sardinha
(2007) [A]
Physical
activity
[A]
133 overweight and obese
community-based females
38.30 (5.80)
[B]
All
IMI
Enjoyment/interest
Effort/importance
Competence
Tension/Pressure
Intervention based on
social cognitive theory
– baseline measures
used in analysis [A]
—
—
Papacharisis, Simou,
& Goudas (2003) [A]
Physical
activity
[A]
643 high school pupils
12.90 (1.20)
[A]
Intention
Attitude
PBC (barriers)
IMI
Enjoyment/interest
Effort/importance
Competence
Tension/Pressure
Correlational – single
wave [A]
—
—
Pihu, Hein, Koka, &
Hagger (in press) [A]
Physical
activity
[A]
399 (276 females, 123 males)
high school pupils
14.70 (1.40)
[A]
Intention
Attitude
PBC
PLOC
Intrinsic,
Identified,
Introjection,
External;
Perceived teacher feedback and learning styles
Correlational,
prospective – 5 weeks
[A]
Self-report –
LTEQ
—
Rentzelas & Hagger
(2008) [B]
Condom
use [B]
84 (69 females, 15 males)
undergraduate and
postgraduate students
22.06 (3.92)
[B]
All
PLOC
Intrinsic,
Identified,
Introjection,
External
Correlational,
prospective – 5 weeks
[A]
Self-report –
2 items
Self-report –
1 item
Sarrazin, Vallerand,
Guillet, Pelletier, &
Cury (2002) [A]
Drop out
from sport
activity
[A]
335 female handball players
14.07 (0.79)
[A]
Intention
SMS
Intrinsic to know,
Intrinsic accomplishment,
Intrinsic stimulation,,
Identified,
Introjection,
External
Amotivation
Correlational,
prospective – 21
months [A]
Dropout
from
handball
programme
—
Shen, McCaughtry, &
Martin (2007) [A]
Physical
activity
[A]
653 high school pupils (335
females, 318 males)
12.4
Range 11-15
[A]
All
PLOC
Intrinsic,
Identified,
Introjection,
External;
IMI
Competence
Correlational,
prospective – 5 weeks
[A]
Self-report -
LTEQ
—
Standage, Duda, &
Ntoumanis (2003) [A]
Physical
activity
[A]
328 high school pupils (138
females, 160 males)
13.56 (0.59)
Range 12-14
[A]
Intention
SMS
Intrinsic (know),
Intrinsic (accomplishment),
Intrinsic (stimulation),
Identified,
Introjection,
External
Amotivation
Correlational – single
wave [A]
—
—

Planned behaviour and self-determination theory 37
Study
Health
Behaviour
Sample(s)1
Mean Age of
Sample (SD)
TPB Constructs
Measured
SDT Constructs Measured
Study Design
Behaviour
Measure
Past
Behaviour
Measure
Thøgerson-Ntoumani
& Ntoumanis (2007)
[A]
Physical
activity
[A]
376 fitness clubs attendees
(246 females, 121 males)
38.7 (10.9)
Range 16-66
[B]
Intention
PLOC (BREQ)
Intrinsic,
Identified,
Introjection,
External
Correlational – single
wave [A]
—
—
Vierling, Standage, &
Treasure (2007) [A]
Physical
activity
[A]
239 elementary school pupils
(119 females, 120 males)
12.11 (1.21)
Range 9.81-
14.41 [A]
Attitude
PLOC
Intrinsic,
Identified,
Introjection,
External;
Perceived autonomy support (adapted WCQ)
Correlational,
prospective – 2 weeks
[A]
Pedometer
step counts
over 2 weeks
—
Vlachopoulos,
Karageorghis, & Terry
(2000) [A]
Sport
participati
on [A]
1145 sports participants;
Sample 1 590 (236 females,
353 males, 1 non-respondent);
Sample 2 555 (250 females,
305 males)
Sample 1:
23.35 (7.54)
Range 18-32;
Sample 2:
23.48 (6.56)
Range 18-30;
[B]
Intention
Attitude
SMS
Intrinsic to know,
Intrinsic accomplishment,
Intrinsic stimulation,,
Identified,
Introjection,
External
Amotivation
Correlational – single
wave [A]
—
Regularity of
sports
participation
– 1 item
Wallhead & Hagger
(2008) [B]
Physical
activity
[A]
189 (95 females, 97 males)
high school pupils; Caucasian
sample 136 (67 females, 69
males); American Indian
sample 56 (28 females, 28
males)
Caucasian:
11.13 (1.25);
American
Indian: 10.47
(0.63) [A]
Intention
Attitude
PBC
Subjective norm
(American
Indian sample
only)
PLOC
Intrinsic,
Identified,
Introjection,
External;
Perceived autonomy support (PASSES)
Correlational, single
wave [A]
—
—
Wells, Thompson, &
Kloeblen-Tarver
(2002)
Breast-
feeding
228 pregnant women from
prenatal clinics
23.00 Range
13-45 [U]
Intention
Intrinsic Motivation
Extrinsic Motivation
Correlational, single
wave [A]
—
5-item
measure of
type and
duration of
breast feeding
Wilson & Rodgers
(2004) [A]
Physical
activity
[A]
232 female university students
and staff
20.86 (2.21)
Range 17-31
[B]
Intention
PLOC (BREQ-2)
Intrinsic,
Identified,
Introjection,
External;
Perceived autonomy support (adapted HCCQ)
Correlational – single
wave [A]
—
—
Wilson, Rodgers,
Blanchard, & Gessell
(2003) [A]
Physical
activity
[A]
53 (44 females, 9 males)
community volunteers
41.75 (10.75)
[B]
Attitude
PLOC (BREQ)
Intrinsic,
Identified,
Introjection,
External
Non-theory based
exercise intervention
[A]
Self-report –
LTEQ
—
Note. [A] = Denotes studies coded for moderators as: published, physical activity behaviours, younger participants, and correlational design; [B] =
Denotes studies coded for moderators as: unpublished, other health behaviours, older participants, and experimental/intervention in design; LTEQ =

Planned behaviour and self-determination theory 38
Table 1 Continues
Leisure time exercise questionnaire (Godin & Shephard, 1985); TPB = Theory of planned behaviour; SDT = Self-determination theory; PLOC =
Perceived locus of causality; PBC = Perceived behavioural control; BREQ = Behavioural regulations in exercise questionnaire (Mullan, Markland, &
Ingledew, 1997); BREQ-2 = Behavioural regulations in exercise questionnaire-2 (Markland & Tobin, 2004); HCCQ = Health care climate
questionnaire (G.C. Williams, Cox, Kouides, & Deci, 1999); IMI = Intrinsic motivation inventory (McAuley, Duncan, & Tammen, 1989); PASSES =
Perceived autonomy support scale for exercise settings (Hagger et al., 2007); PECCS = Physical education class climate scale (Biddle et al., 1995);
SRQ = Self-regulation questionnaire (Goudas, Biddle, & Fox, 1994); LCQ = Learning climate questionnaire (G. C. Williams & Deci, 1996); SMS =
Sport Motivation Scale (Briere, Vallerand, Blais, & Pelletier, 1995); PASE = Physical activity scale for the elderly (Washburn, Smith, Jette, & Janney,
1993); WCQ = Work Climate Questionnaire (Baard, Deci, & Ryan, 2004).

Planned behaviour and self-determination theory 39
Table 2 Continues
Table 2.
Results of Meta-Analysis of Theory of Planned Behaviour and Self-Determination Theory
Components
Effect
k
N
r+a
r++b
CI95
CI90
SD
SE
NFS
Varc
LB
UB
LB
UB
Self-determined motivation—
Perceived autonomy support
18
4036
.32
.38
.32
.44
.20
.56
.11
.03
418
31.62
Self-determined motivation—
Intention
38
10784
.44
.52
.46
.57
.25
.79
.16
.03
6325
12.64
Self-determined motivation—
Attitude
28
7296
.45
.54
.44
.64
.13
.95
.25
.05
3499
6.33
Self-determined motivation—
Subjective norm
18
4489
.19
.24
.15
.33
-.05
.54
.18
.05
170
17.21
Self-determined motivation—
PBC
22
5835
.37
.46
.35
.57
.05
.87
.25
.05
1215
7.34
Self-determined motivation—
Behaviour
28
5505
.30
.37
.30
.45
.08
.67
.18
.04
1765
18.32
Self-determined motivation—
Past behaviour
18
4041
.28
.34
.26
.43
.06
.62
.17
.04
611
17.65
Perceived autonomy support—
Intention
19
4139
.24
.28
.25
.32
.05
.51
.14
04
492
23.09
Perceived autonomy support—
Attitude
15
2715
.29
.32
.24
.41
.10
.55
.14
.04
327
24.02
Perceived autonomy support—
Subjective norm
11
1862
.21
.27
.16
.37
.03
.50
.14
.05
65
31.98
Perceived autonomy support—
PBC
13
2397
.15
.19
.11
.26
.01
.37
.11
.04
67
37.67
Perceived autonomy support—
Behaviour
14
2636
.20
.25
.17
.33
.05
.44
.11
.04
110
34.69
Perceived autonomy support—
Past behaviour
11
2021
.23
.24
.14
.33
.01
.47
.14
.05
76
21.36
Intention—Attitude
26
6662
.59
.70
.64
.75
.50
.90
.12
.03
5673
21.11
Intention—Subjective norm
21
5005
.33
.43
.33
.52
.09
.76
.20
.05
1184
14.14
Intention—PBC
24
5708
.51
.62
.53
.71
.27
.97
.22
.05
3011
9.04
Intention—Behaviour
27
5594
.52
.62
.54
.70
.28
.95
.21
.04
5347
10.72
Intention—Past behaviour
19
4171
.48
.57
.45
.70
.12
1.02
.27
.06
2413
5.99
Attitude—Subjective norm
20
4831
.32
.42
.31
.53
.04
.79
.23
.05
972
11.14
Attitude—PBC
23
5534
.45
.54
.47
.61
.29
.79
.15
.04
2295
16.48

Planned behaviour and self-determination theory 40
Effect
k
N
r+a
r++b
CI95
CI90
SD
SE
NFS
Varc
LB
UB
LB
UB
Attitude—Behaviour
24
4840
.37
.45
.38
.52
.21
.69
.15
.03
2222
22.08
Attitude—Past behaviour
18
3956
.34
.42
.30
.54
-.01
.84
.26
.06
895
8.54
Subjective norm—PBC
20
4831
.27
.37
.26
.47
.01
.72
.22
.05
598
13.68
Subjective norm—Behaviour
18
3610
.19
.25
.17
.33
.01
.49
.15
.04
342
28.86
Subjective norm—Past
behaviour
15
2739
.27
.35
.25
.45
.07
.63
.17
.05
265
24.22
PBC—Behaviour
22
4487
.30
.38
.29
.48
.03
.73
.21
.05
1248
13.61
PBC—Past behaviour
16
2907
.29
.37
.26
.48
.03
.72
.21
.06
493
15.14
Behaviour-past behaviour
17
3081
.57
.73
.61
.84
.35
1.10
.23
.06
2360
10.79
Note. aAveraged correlation corrected for sampling error only; bAveraged correlation
corrected for sampling error and measurement error; cVariance accounted for by statistical
artifacts of sampling and measurement error. PBC = Perceived behavioural control; k =
Number of effect sizes contributing to averaged corrected correlation from the meta-analysis;
N = total sample size across studies contributing to correlation; CI95 = 95% confidence
intervals for averaged correlation corrected for sampling error only; CI90 = 90% confidence
interval for averaged correlation corrected for measurement error; LB = Lower bound of
confidence/credibility interval; UB = Upper bound of confidence/credibility interval; SD =
Standard deviation of averaged correlation corrected for sampling and measurement error; SE
= Standard error of averaged correlation corrected for sampling and measurement error; NFS =
Fail safe N.

Planned behaviour and self-determination theory 41
Table 3 continues
Table 3.
Results of Moderator Analyses for Effects of the Theory of Planned Behaviour and Self-Determination Theory Components
Moderator
Effect
k
N
r+a
r++b
CI95
CI90
SD
SE
Varc
LB
UB
LB
UB
Age
Self-determined
motivation—PBC
15
(7)
4129
(1228)
.48
(.33)
.59
(.39)
.51
(.29)
.67
(.48)
.34
(.22)
.83
(.54)
.14
(.09)
.04
(.05)
16.13
(39.92)
Behaviour
type
Attitude—PBC
20
(3)
5047
(487)
.47
(.25)
.57
(.30)
.50
(.22)
.63
(.37)
.33
(.29)
.80
(.29)
.14
(.00)
.04
(.04)
17.98
(100.00)
Attitude—Past behaviour
15
(3)
3534
(422)
.31
(.59)
.39
(.70)
.25
(.65)
.52
(.74)
-.02
(.70)
.80
(.70)
.25
(.00)
.07
(.02)
8.80
(100.00)
Intention—Past behaviour
16
(3)
3684
(487)
.45
(.71)
.52
(.89)
.39
(.87)
.65
(.91)
.08
(.89)
.95
(.89)
.26
(.00)
.07
(.01)
5.37
(100.00)
PBC—Behaviour
19
(3)
4000
(487)
.32
(.10)
.41
(.13)
.31
(.01)
.51
(.24)
.07
(.05)
.75
(.20)
.20
(.05)
.05
(.06)
14.09
(81.84)

Planned behaviour and self-determination theory 42
Moderator
Effect
k
N
r+a
r++b
CI95
CI90
SD
SE
Varc
LB
UB
LB
UB
Study
design
Self-determined
motivation—Intention
34
(4)
10385
(399)
.45
(.31)
.53
(.32)
.47
(.24)
.58
(.39)
.26
(.32)
.79
(.32)
.16
(.00)
.03
(.03)
11.85
(100.00)
Note. Statistics for younger participants, physical activity behaviours, and correlational studies are shown without parentheses and statistics for
older participants, other health behaviours, and experimental/intervention studies are shown in parentheses. aAveraged correlation corrected for
sampling error only; bAveraged correlation corrected for sampling error and measurement error; cVariance accounted for by statistical artifacts of
sampling and measurement error. PBC = Perceived behavioural control; k = Number of effect sizes contributing to averaged corrected correlation
from the meta-analysis; N = total sample size across studies contributing to correlation; CI95 = 95% confidence intervals for averaged correlation
corrected for sampling error only; CI90 = 90% confidence interval for averaged correlation corrected for measurement error; LB = Lower bound
of confidence/credibility interval; UB = Upper bound of confidence/credibility interval; SD = Standard deviation of averaged correlation
corrected for sampling and measurement error; SE = Standard error of averaged correlation corrected for sampling and measurement error.

Planned behaviour and self-determination theory 43
Figure caption
Figure 1. Meta-analytic path analysis of the proposed motivational sequence arising from the
integration of the Theory of Planned Behaviour and Self-Determination Theory.
Note.
Coefficients are standardized regression coefficients. Only statistically significant paths
shown.
* p < .05 ** p < .01.

Planned behaviour and self-determination theory 44
Intention
Subjective
Norm
Perceived
Behavioral
Control
Attitude
Behavior
.27**
.06**
.31**
.37**
.44**
.38**
.23**
Perceived
Autonomy
Support
Self-
Determined
Motivation
Past
Behaviour
.55**
.27**
.24**
.27**
.14**
.31**
.24**
.29**
.10**