Content uploaded by Martin Hagger
Author content
All content in this area was uploaded by Martin Hagger on Mar 31, 2016
Content may be subject to copyright.
284 / Chatzisarantis, Hagger, Biddle, Smith, and Wang
284
JOURNAL OF SPORT & EXERCISE PSYCHOLOGY, 2003, 25, 284-306
© 2003 Human Kinetics Publishers, Inc.
A Meta-Analysis of Perceived Locus of Causality
in Exercise, Sport, and Physical Education Contexts
Nikos L.D. Chatzisarantis1, Martin S. Hagger2, Stuart J.H. Biddle3,
Brett Smith4, and John C.K. Wang5
1,4University of Exeter, 2University of Essex, 3Loughborough University,
5Nanyang Technical University
The present article conducts a meta-analytic review of the research adopting
the perceived locus of causality in the contexts of sport, exercise, and physical
education. A literature search of published articles identified three main re-
search foci: (a) the development of instruments that assess perceived locus of
causality; (b) examination of the construct validity of perceived locus of cau-
sality by investigating the relevance of the self-determination continuum as
well as by using antecedents (e.g., perceived competence) and outcomes (e.g.,
intentions); and (c) integration of Nicholls’ (1984) concepts of task and ego
orientation with perceived locus of causality. A meta-analysis using 21 pub-
lished articles supported the existence of a self-determination continuum from
external regulation to introjection and identification. In addition, path analy-
sis of corrected effect sizes supported the mediating effects of perceived locus
of causality on the relationship between perceived competence and intentions.
Results are discussed with reference to the assumptions of self-determination
theory, Vallerand’s (1997) hierarchical model of intrinsic/extrinsic motiva-
tion, and theories of behavioral intentions.
Key Words: self-determination theory, perceived competence, behavioral
intentions
Understanding adherence to health-related behaviors is an important avenue
for scientific inquiry. Knowledge of the fundamental processes and mechanisms
of human behavior can inform practice about how to promote human motivation.
Deci and Ryan’s (1985) self-determination theory has become increasingly popu-
lar in studies of human motivation (Deci, Koestner, & Ryan, 1999a).
1Dept. of Sport & Health Sciences, Univ. of Exeter, St Luke’s Campus, Devon EX1
2LU, U.K.; 2Dept. of Psychology, Univ. of Essex, Wivenhoe Park, Colchester, Essex, CO4
3SQ, U.K.; 3Dept. of P.E., Sports Science & Recreation Management, Loughborough Univ.,
Loughborough, Leics., LE11 3TU, U.K.; 4Qualitative Research Unit, University of Exeter;
5Physical Education and Sports Sciences, Nanyang Technical University, Singapore 637616.
A Meta-Analysis of PLOC / 285
Self-Determination Theory
Its Approach to Human Behavior
Self-determination theory postulates that intentional human behavior can be
described, in a parsimonious way, through two processes of intrinsic motivation
and internalization. Intrinsic motivation refers to “the doing of an activity for its
inherent satisfactions rather than for some separable consequences” (Ryan & Deci,
2000, p. 56). Cognitive evaluation is a sub-theory of self-determination theory that
attempts to understand factors that facilitate and undermine intrinsic motivation. It
has been postulated that intrinsic motivation is engendered when people are in
conditions that support three innate psychological needs: the need for self-deter-
mination, competence, and relatedness (Ryan & Deci, 2000). Self-determination
refers to the need to initiate and regulate one’s own actions. Competence refers to
the need to produce behavioral outcomes and to understand production of these
behavioral outcomes. Relatedness refers to the need to have satisfactory relation-
ships with others and with the social order in general (Deci & Ryan, 1990).
In a meta-analysis of experimental studies dealing with intrinsic motivation,
Deci et al. (1999a) established a relationship between experimental conditions in-
fluencing psychological needs and intrinsic motivation. In the experiments they
meta-analyzed, they assessed intrinsic motivation after exposing individuals to
conditions that either frustrated or satisfied psychological needs. The psychologi-
cal need for self-determination was manipulated by exposing individuals to condi-
tions of either choice or no choice. In addition, the need for competence was
manipulated by giving positive or negative feedback.
Following exposure to such conditions, participants’ levels of intrinsic moti-
vation were assessed in two ways. First, engagement in a target task during which
individuals were allowed to engage in alternative interesting tasks (free-choice
period) was used to represent a behavioral indicator of intrinsic motivation. Sec-
ond, a self-report measure of interest derived from the task chosen during the free-
choice period was used as a more covert measure of intrinsic motivation. Deci et
al. (1999a) reported that conditions which frustrated psychological needs under-
mined self-reported interest and overt involvement with the target task when com-
pared to conditions that facilitated the satisfaction of such needs. In addition, there
is evidence that intrinsically motivated behaviors are intentional. Chaiken (1980)
showed that people are more likely to express intentions to search information
about a topic when the topic is personally interesting to them vs. when it is not
(Chatzisarantis, Hagger, Biddle, & Karageorghis, 2002).
Although the concept of intrinsic motivation has attracted a great deal of
scientific interest and debate (see Eisenberg, Pierce, & Cameron, 2000), behav-
ioral regulation through intrinsic motivation is not the only type of social behavior
that individuals can engage in. For this reason, Deci and Ryan (1985) proposed an
organismic integration theory, which is a second sub-theory of self-determination
theory, to explain the process of internalization. Internalization is the process through
which individuals take in a value or regulation and progressively transform it so
that the regulation emanates from their own sense of self.
A model describing internalization and human motivation from an organis-
mic integration theory perspective is shown in Figure 1a. On the left side is
amotivation, referring to a person’s lack of intentionality and sense of personal
286 / Chatzisarantis, Hagger, Biddle, Smith, and Wang
Figure 1 — (a and b). Simplex order structure of PLOC. Integration not featured because very few studies included measures of
integration in PLOC scale. Note: Identified regulation is a dependent variable, therefore covariances between identification and
intrinsic motivation cannot be estimated. For this reason, paths are specified from identification to intrinsic motivation.
a)
b)
A Meta-Analysis of PLOC / 287
causation. Perceived incompetence and beliefs that behavior cannot reliably lead
to desired outcomes can precipitate amotivation. To the right of amotivation on the
self-determination theory continuum are four forms of extrinsic motivation: exter-
nal regulation, introjection, identification, and integration. Each form of regula-
tion reflects the achievement of outcomes that are separate from the behavior,
which is why they do not represent intrinsic motivation. Even so, they do reflect
different degrees of internalization.
Internalization is initialized by significant others to whom one feels attached
or related (Ryan & Deci, 2000). These externally prompted behaviors are repre-
sented by external regulation. Introjection lies next to external regulation and re-
fers to a behavior performed in order to avoid a pressuring emotion of guilt or
shame. External regulation and introjection describe less internalized and more
controlling forms of behavior because they refer to behaviors performed under
internal (i.e., introjection) or external pressure (i.e., external regulation). A less
controlling and more autonomous form of behavior is described by identification,
a behavior that is performed because the individual values it. During identifica-
tion, individuals accept and endorse the value of behavior as a reason for action,
thus, identified behavior reflects higher degrees of internalization.
The most autonomous and least controlling form of behavior is integrated
regulation, which refers to identifications that one brings into congruence with
other behaviors and roles he or she enacts in life. This definition of integrated
regulation presupposes that identification is a less autonomous form of behavioral
regulation than integration, because regulation through identification may conflict
with preexisting values and behaviors. In their study, Deci, Eghrari, Patrick, and
Leone (1994) established a relationship between experimental conditions that in-
fluence psychological needs and internalization of an initially boring activity. In
this experiment, the provision of a rationale, acknowledgement of tension and
pressure associated with involvement in a boring activity, provision of choice, and
use of nonpressuring language (e.g., “may” or “could” vs. “should” or “must”)
were all found to facilitate internalization.
Ryan and Connell (1989) developed an instrument, in an educational con-
text, termed the Perceived Locus of Causality scale (PLOC). The PLOC measures
external regulation, introjection, identification, and intrinsic motivation. These
dimensions of the PLOC scale were also found to conform to a simplex-ordered
structure. A simplex-ordered structure is evident when correlations between adja-
cent types of behavior (e.g., external regulation and introjection) are higher than
correlations between dimensions that lie further apart (e.g., external regulation
and identification). Furthermore, simplex-ordered correlation matrices indicate the
presence of a continuum, which Deci and Ryan (1990) describe as a developmen-
tal continuum of self-determination. However, Ryan and Deci (2000) also suggest
that support of a continuum does not preclude the possibility for individuals to
internalize a new behavior at any point along this continuum, depending on prior
experience and current situational factors.
Applications in Sport, Exercise, and Physical Education
The concept of PLOC provides a very comprehensive view of human moti-
vation. As Ryan and Deci (2000) put it, “even a brief reflection suggests that mo-
tivation is hardly a unitary phenomenon” (p. 54), yet many theories of intentional
behavior still define and operationalize motivation as a unitary phenomenon
288 / Chatzisarantis, Hagger, Biddle, Smith, and Wang
(Bagozzi & Kimmel, 1995). In sport, exercise, and physical education contexts,
research has dealt with the development and examination of construct validity of
instruments that measure intrinsic motivation and internalization. A number of
studies have also examined relationships between measures of PLOC and Nicholls’
(1984) achievement goals theory (Biddle, Soos, & Chatzisarantis, 1999; Goudas,
Biddle, & Fox, 1994; Ntoumanis 2001b; Wang, Chatzisarantis, Spray, & Biddle,
2002).
With regard to measurement issues, there is a French version of the Sport
Motivation Scale (SMS), developed by Briere, Vallerand, Blais, and Pelletier
(1995) and translated into English by Pelletier, Fortier, Vallerand, Tuson, and
Blais (1995); translated into Bulgarian by Chantal, Guay, and Martinova (1996);
and translated into Greek by Georgiadis, Biddle, and Chatzisarantis (2001). The
Sport Motivation Scale assesses amotivation, external regulation, introjection, iden-
tification, and three types of intrinsic motivation: to know (e.g., “for the pleasure it
gives me to know more about the sport that I practice”); to accomplish (e.g., “be-
cause I feel a lot of personal satisfaction while mastering certain difficult move-
ments”); and to experience stimulation (e.g., “for the intense emotions”). In the
context of physical activity is a Behavioral Regulation in Exercise Questionnaire
(BREQ) developed by Mullen, Markland, and Ingledew (1997), and the Exercise
Motivation Scale (EMS) developed by Li (1999). Finally, Goudas et al. (1994)
adapted Ryan and Connell’s (1989) Self-Regulation Questionnaire as well as
Vallerand, Pelletier, and Blais et al.’s (1992) Academic Motivation Scale in the
physical education context. This PLOC scale for physical education measures ex-
ternal regulation, introjection, identification, and intrinsic motivation, but not inte-
grated regulation.
Despite differences in measurement, correlations between different types of
behavior, described by PLOC, have been shown to conform to a simplex-ordered
structure. Therefore, studies do support the hypothesized self-determination con-
tinuum. However, studies have predominantly used a visual inspection of correla-
tions (observed, or correlations corrected for attenuation) in inferring the presence
of the continuum. An observed correlation matrix is simplex-ordered when a model
specifying linear dependencies between adjacent dimensions only, and not between
nonadjacent dimensions, explains the observed correlation matrix satisfactorily
(Joreskog, 1970). Tests of this kind can be conducted with structural equation
modeling techniques, on a LISREL or EQS interface. In the sport context, only
two studies, Li and Harmer (1996) and Li (1999), confirmed the simplex pattern of
the Sport Motivation Scale and the Exercise Motivation Scale through structural
equation modeling techniques. Thus it is not yet known whether a simplex model
can represent results (i.e., correlations) from studies using PLOC in the contexts of
sport, exercise, and physical education.
Research on sport has also explored associations between PLOC and other
psychological variables in an attempt to examine the construct validity of PLOC.
Briere et al. (1995) and Pelletier et al. (1995) found positive correlations between
more autonomous forms of behavior (identification and intrinsic motivation) with
hypothesized determinant (e.g., perceived autonomy, competence) and outcome
variables (e.g., effort, intentions). Moreover, in the context of physical education,
Goudas et al. (1994) found associations between PLOC and Nicholls’ (1984) con-
cepts of task and ego orientation (see also Ntoumanis, 2001b). Two studies pointed
A Meta-Analysis of PLOC / 289
to positive relationships between perceived relatedness, autonomy, and PLOC (Li,
1999; Ntoumanis, 2001a). This empirical evidence supports the construct validity
of PLOC because PLOC is correlated with hypothesized determinant and outcome
variables (Briere et al., 1995; Goudas et al., 1994; Pelletier et al., 1995). However,
correlational studies also raise an important question about the role of PLOC in the
prediction and explanation of physical activity.
In the physical education context, research has addressed this question but
yielded equivocal results. Using path analysis, Goudas et al. (1994) found perceived
competence to mediate the effects of PLOC and of task and ego orientations on inten-
tions. However, Goudas, Biddle, and Underwood (1995) also point out that per-
ceived competence covaries with PLOC in predicting intentions. Yet, in another
study Biddle et al. (1999) found PLOC to mediate effects of perceived compe-
tence and of task and ego orientation on intentions (Ntoumanis 2001a). Hence the
role of PLOC in predicting and explaining physical activity is still an open question.
Research Hypotheses
The present study examines the adequacy of three models in explaining re-
sults from studies using PLOC and other psychological antecedents and outcomes
of physical activity. The adequacy of the models is tested through a combination
of meta-analytic and path-analytic techniques (Hunter & Schmidt, 1990; Tett &
Meyer, 1993). Meta-analytic applications permit researchers to make inferences
about a relationship after correcting correlation coefficients and variance of corre-
lations for statistical artifacts of sampling error and measurement error (reliabil-
ity). The logic of meta-analysis is that, in many instances, equivocal results are
published because of statistical artifacts (sampling and measurement error), and
controlling for such statistical artifacts permits more accurate inferences about a
set of relationships.
The present study will correct two correlation matrices for sampling error
and measurement error by using a model of random effect sizes (Schmidt & Hunter,
1999). The first meta-analysis will correct correlations and variance of correla-
tions between dimensions of PLOC. Whether or not the corrected correlations
between dimensions of PLOC conform to a simplex-ordered structure will then be
examined through path analysis. Given that in the exercise and sport psychology
literature, intrinsic motivation is operationally defined as either one- or three-
dimensional, the adequacy of a one-dimensional and a three-dimensional model in
explaining correlation matrices will be examined (see Figure 1).
The second meta-analysis will correct correlation coefficients and variance
of correlations between PLOC, perceived competence, and intentions. The extent
to which PLOC mediates the effects of perceived competence on intentions will
then be examined using path analysis (Figure 2). The present study will not meta-
analyze the relationships between perceived autonomy, relatedness, and PLOC,
since only a few studies employed measures of relatedness and autonomy. Be-
cause contemporary research has treated perceived competence as an antecedent
of PLOC and intentions as an outcome variable (Briere et al., 1995; Pellettier et
al., 1995; Vallerand, 1997), it is hypothesized that PLOC will mediate the effects
of perceived competence on intentions. This final hypothesis reflects our interpre-
tation of Deci and Ryan’s (1985) self-determination theory.
290 / Chatzisarantis, Hagger, Biddle, Smith, and Wang
Figure 2 — A model showing antecedents and outcomes of PLOC.
A Meta-Analysis of PLOC / 291
Method
Selection of Studies
We conducted an electronic CD-ROM search using the following key words:
self-determination theory, organismic-integration theory, perceived locus of cau-
sality, self-determination continuum, autonomy continuum, physical exercise, sport,
athletics, physical education classes, school, and classroom. We used the follow-
ing databases: Sport Discus (1974–2001), Psychlit (1974–2003), ERIC and Brit-
ish Educational Index (1974 to 2003), Australian Educational Index (1978 to 2003)
and Canadian Educational Index (1976 to 2003), and the Index to Theses. We used
Ryan and Deci’s (2000) review and the meta-analysis conducted by Deci, Koestner,
and Ryan (1999a, 1999b) to locate other articles that might not have been included
in the electronic databases.
From the pool of articles using the self-determination theory, studies were
rejected on the basis of three criteria: (a) did not include statistical data; (b) did not
use measures of PLOC; and (c) did not use at least two dimensions of PLOC (see
Markland & Hardy, 1997). Based on these criteria, we subjected the results from
21 articles to statistical analysis. The studies are presented in Table 1. Meta-ana-
lytic correlations between dimensions of PLOC and competence, intentions, and
PLOC were estimated on the basis of cross-sectional observations. Note that meta-
analyses which include a small number of studies are as important as those which
include large number of studies, given that they can be used to conduct second-
order meta-analysis (Hunter & Schmidt, 1990; Schmidt, 1996). However, when
meta-analyzing a small number of studies, it is important to control for inflated
Type I error rates that the Hunter and Schmidt meta-analysis procedure produces
(Field, 2001). In the present meta-analysis, the predicted Type I error rates of rela-
tionships that are hypothesized to be significant range from .07 for the external
regulation/introjection relationship to .15 for the introjection/intrinsic-motivation-
to-know relationship. To reduce Type I error rates, we accepted a correlation as sig-
nificant when it was three times as large as its respective standard error (Field, 2001).
Statistical Analysis
The present meta-analysis gathered information from each study in terms of
sample size, reliability, age, gender, purpose, context, design, measures used in the
study, and type of PLOC measure. Based on sample and reliability information,
we corrected the correlation coefficients of each study and variance of correlations
across studies for sampling error and measurement error. Correlations were not
corrected for restriction or enhancement in range because there is not a normative
study reporting the population standard deviations of PLOC (Hunter & Schmidt,
1990). We attempted to retrieve as much information as possible from authors (via
email) and/or to estimate correlation coefficients from statistics as reported in the
articles. Not all the required statistical information could be obtained, and there-
fore results of the present meta-analyses were based on distributions of correlation
coefficients and of reliability information. According to Hunter, Schmidt, and Jack-
son (1982), such meta-analysis is as accurate as meta-analysis using complete in-
formation.
292 / Chatzisarantis, Hagger, Biddle, Smith, and Wang
Table 1 Characteristics of Studies Using Measures of PLOC
Studies Sample
Age of participants size Purpose Context Design Measures used Measures of PLOC
Biddle et al. (1999) N = 723 Look at physical activity Sport & Cross- Task & ego orientation, Adaptation of Ryan &
12–16 yrs intentions from a goals & physical sectional perceived competence Connell’s PLOC scale
self-determination theory education PLOC scale, intentions
perspective
Briere et al. (1995) 122 M Convergent validity of Sport Cross- Sport motivation scale French version of
Study 1 73 F French version of sectional Sport Motivation Scale
18.46 yrs sport motivation scale
Chantal et al. (1996) 63 M Examined elite athletes’ Com- Cross- Sport motivation scale, Bulgarian version of
19.5 yrs 35 F levels of self-determination petitive sectional titles & medals of athletes Pelletier et al.’s (1995)
sport sport motivation scale
Chatzisarantis et al. 21 M Relationship between School Prospec- Intentions, physical Mullen et al.’s (1997)
(1997) 79 F self-determination theory, settings tive activity during leisure behavioral regulation
13.5 yrs intentions, & generality & leisure time scale
of behavioral change
Chatzisarantis & 51 M Self-determination theory Leisure Cross- Phys. activity, intentions, Adaptation of Ryan &
Biddle (1998) 50 F & intention formation time sectional PLOC for phys. activity Connell’s PLOC scale
40 yrs attitudes, subjective norms,
intentions, perceived control
Chatzisarantis et al. 78 M Prediction of physical Leisure Prospec- Attitudes, control, effort, Mullen et al.’s (1997)
(2002) 62 F activity tive past behavior, intentions, behavioral regulations
13.5 yrs & PLOC for phys. activity scale
A Meta-Analysis of PLOC / 293
Georgiadis et al. N = 350 Prediction of global & Leisure Cross- Goal orientations, PLOC, Greek version of
(2001) physical self-worth sectional global self-worth & Pelletier et al. ‘s (1995)
30.8 yrs physical self-worth sport motivation scale
Goudas et al. (1994) 39 M Looking at phys. activity Physical Cross- Goal orientations, PLOC, Adaptation of Ryan &
12–14 yrs 46 F intentions from a goals education sectional perceived competence, Connell’s PLOC scale
& self-determination intrinsic interest, intention,
theory perspective activity preference
Goudas et al. (1995) 23 M Examination of Physical Prospec- PLOC, perceived compe- Vallerand et al.’s
20–25 yrs 17 F determinants of education tive tence, intrinsic motivation, academic motivation
intrinsic motivation intention, performance scale
Hagger et al. (2002) 551 M Links between PLOC Leisure Cross- PLOC for leisure time, Adaptation of Ryan
12–14 yrs 537 F & theory of planned time sectional & theory of planned & Connell’s (1989)
behavior behavior PLOC scale
Hamer et al. (2002) 147 M Prediction of running Amateur Prospec- PLOC, running addiction Mullan et al.’s (1997)
35.8 yrs 41 F addiction sport tive scale behavioral regulations
scale
Li (1999) Study 2 N = 371 Develpmnt. & validity of an Leisure Cross- Exercise motivation scale Exercise motivation scale
College age exercise motivation scale time sectional
Li (1999) Study 3 205 M Examined hierarchical & Leisure Cross- Exercise motivation scale, Exercise motivation scale
21.49 yrs 393 F simplex structure of the time sectional perceptions of exercise
exercise motivation scale competence, exercise auto-
nomy, relatedness, exerc.
interest, exercise effort
(continued)
294 / Chatzisarantis, Hagger, Biddle, Smith, and Wang
Li & Harmer (1996) 442 M Examined the simplex Sport Cross- Sport motivation scale Pelletier et al.’s (1995)
20.93 yrs 415 F structure of Pelletier et al.’s sectional sport motivation scale
(1995) sport motivation
scale across gender
Mullen & Markland 158 M Relationship between PLOC Leisure Cross- PLOC scale for physical Mullan et al.’s (1997)
(1997) 157 F scale for exercise & stages time sectional activity, stages of change behavioral regulations
37.4 yrs of change scale
Ntoumanis (2001a) N = 253 Relationship between PLOC Phys Ed Cross- PLOC, competence, Adaptation of Ryan
15.7 yrs scale, psychological needs & leisure sectional relatedness, autonomy, & Connell’s (1989)
& intentions time & intentions PLOC scale
Ntoumanis (2001b) N = 268 Relationship between goal Sport Cross- Goal orientations, Pelletier et al.’s (1995)
20.4 yrs orientations & PLOC sectional PLOC, competence sport motivation scale
Pelletier et al. 319 M Translation of sport motiv. Sport Cross- Sport motivation scale Briere et al.’s (1995)
(1995) Study 1 274 F scale from French to Engl. sectional French version of
19.2 yrs Construct validity of sport motivation scale
sport motivation scale
Standage, Duda, N = 328 Prediction of physical Phys ed Cross- Perceived climate in Pelletier et al.’s (1995)
Ntoumanis (2003) activity intentions & leisure sectional phys educ, comptence, sport motivation scale
13.56 yrs time relatedness, autonoomy
Table 1 (Continued)
Studies Sample
Age of participants size Purpose Context Design Measures used Measures of PLOC
A Meta-Analysis of PLOC / 295
Standage, Treasure, N = 439 Validity of situational Sport Cross- Situational motivation Situational motivation
et al. (2003) Study 1 motivation scale sectional scale scale
16.13 yrs
Standage, Treasure, 182 M Validity of situational Physical Cross- Situational motivation Situational motivation
et al. (2003) Study 1 136 F motivation scale education sectional scale scale
13.2 yrs
Standage, Treasure, 99 M Validity of situational College Cross- Situational motivation Situational motivation
et al. (2003) Study 1 122 F motivation scale phys. act. sectional scale scale
20.82 yrs course
Standage, Treasure, 1,008 F Validity of situational Physical Experi- Situational motivation Situational motivation
et al. (2003) Study 2 motivation scale education mental scale scale
12–14 yrs
Vallerand & Losier N = 77 Self-determination Sport Prospec- Sportsmanship Briere et al.’s (1995)
(1994) gender not & sportsmanship tive orientations French version of
15.8 yrs reported sport motivation scale
Wang et al. (2002) 427 M Goal & motivational Phys ed Cross- PLOC for PE, beliefs re. Goudas et al. PLOC
12.71 yrs 391 F profiles in the context & leisure sectional sports ability, participation for physical education
of physical education time in leisure, phys. activity, used by Goudas et al.
task & ego orientations, (1994)
perceived competence
Note: None of the longitudinal studies assessed intentions, PLOC constructs, and/or competence longitudinally.
296 / Chatzisarantis, Hagger, Biddle, Smith, and Wang
Based on statistical information gathered from the studies, we estimated four
parameters: (a) a correlation coefficient corrected for sampling error; (b) a correla-
tion corrected both for sampling error and measurement error; (c) a standard error,
which reflects variance of correlations corrected for sampling error (Hunter &
Schmidt, 1990); and (d) a standard deviation reflecting variance of correlations,
corrected for sampling error and measurement error (Whitener, 1992).
The corrected correlations, standard errors, and standard deviations can be
used to calculate both confidence and credibility intervals. Confidence intervals can
be constructed by using standard errors around corresponding correlations that are
corrected for sampling error (Whitener, 1992). Large standard errors and confi-
dence intervals suggest that the power of the meta-analysis is low, perhaps owing
to inclusion of a small number of studies. Standard errors and confidence intervals
are therefore estimates of the accuracy of a correlation coefficient. A credibility
interval can be constructed by using standard deviations around the correlations
corrected for sampling error and measurement error. Credibility intervals indicate
the extent to which an association (e.g., a relationship between external regulation
and introjection) is consistent across studies. If artifacts account for less than 75%
of the uncorrected variance (Hunter et al., 1982), the credibility intervals are con-
sidered large and the results across studies are inconsistent. If, by contrast, arti-
facts account for more than the 75% of the uncorrected variance, the credibility
intervals are considered small and the results are consistent across studies.
In the case of large credibility intervals, researchers can use a moderator
analysis to investigate whether study characteristics might explain variance in study
results. A moderator analysis involves the classification of studies into groups on
the basis of some criteria (study characteristics) and then running separate meta-
analyses for each group of studies. With regard to applications of PLOC, reason-
able classification criteria can be the context that instruments of PLOC are referring
to (e.g., sport, physical education, or exercise) or gender (Chantal et al., 1996).
Finally, the present meta-analysis examined the adequacy of the models shown in
Figures 1 and 2 in terms of how they explain the corrected correlation matrices
through path analysis by following Cudeck’s (1989) and Tett and Meyer’s (1993)
recommendations.
Results
Effect Sizes and Consistency of Study Results
Table 2 presents correlation coefficients between dimensions of PLOC, cor-
rected for sampling error and measurement error, as well as standard errors and
standard deviations of these correlation coefficients. The magnitude of the corre-
lations suggests the presence of a continuum, given that the correlation coeffi-
cients between the adjacent factors (e.g., external regulation and introjection) are
greater than those between the factors lying further apart on the hypothesized con-
tinuum (e.g., external regulation and identification). The relatively low standard
error suggests that, in spite of the small number of studies, the correlations are
relatively accurate. The magnitude of standard errors indicates the extent of sec-
ond-order sampling error (Hunter & Schmidt, 1990). It also supports the power of
the present meta-analysis to detect significant relationships between adjacent
dimensions of PLOC at the .001 alpha levels. None of the correlations between
A Meta-Analysis of PLOC / 297
adjacent dimensions of PLOC were statistically insignificant. The relatively large
standard deviations of the correlation coefficients, however, do indicate the pres-
ence of moderating variables.
To examine whether context explains variability in study results, we split
the studies in terms of the context that PLOC instruments referred to (education
vs. sport vs. leisure physical activity) and performed three meta-analyses for each
group of studies. We classified the studies in terms of context on the basis of the
PLOC constructs employed, since primary studies used different PLOC constructs
depending on the context being examined. In comparison to single-group meta-
analysis, the moderator analysis did not reduce the variability in study results, nor
did it point out statistically significant differences between correlations after ap-
plying Hunter and Schmidt’s (1990) z-tests that account for second-order sam-
pling error. However, the power of this moderator meta-analysis is low because
the predicted Type II error rates are estimated to be .50 at the .05 alpha level
(Field, 2001). Given the low power of moderator analysis for detecting differ-
ences between correlations at the .05 alpha level, and even less power at the .001
alpha level, the conclusion that context does not explain variability in study re-
sults must be treated with caution.
The Simplex-Ordered Structure of PLOC
Examination of the hypothesized simplex-ordered structure of correlation
coefficients, corrected for sampling and measurement error, was conducted through
path analysis. Cudeck (1989) suggested that path analysis could be performed on
correlation matrices only if the hypothesized models are not scale-invariant. A
model is scale-invariant when it does not constrain parameter estimates at a fixed
non-zero value. When a model constrains parameters at some non-zero value, it is
not scale-invariant and path analysis on correlation matrices is inappropriate. In
the present study, the hypothesized model is scale-invariant because parameters
are not fixed at non-zero values. For example, the hypothesized model required
path coefficients between the adjacent dimensions of the PLOC to be free, and
path coefficients between nonadjacent dimensions to be fixed to zero.
Figure 1a and 1b displays parameter estimates of two models that are hy-
pothesized to explain correlations corrected for sampling error and measurement
error. Through the model shown in Figure 1a, we can examine the extent to which
a correlation matrix involving only one dimension of intrinsic motivation to expe-
rience stimulation is simplex-ordered. The comparative fit index of this model
was .98, which satisfies recent criteria for good fit (Hu & Bentler, 1999).
With regard to the model’s parameters, amotivation and identification were
positively associated with external regulation and intrinsic motivation, respec-
tively. Moreover, path coefficients supported a positive indirect effect of external
regulation on identification via introjection (indirect effect = .27). However, the
Lagrange Multiplier Test did reveal improvements in model fit if the paths from
amotivation to identification and from external regulation to intrinsic motivation
were released. Estimation of these paths improved model fit (chi-square change =
170.75 p < .001), and negative relationships were found between amotivation and
identification (–.24) and between external regulation and intrinsic motivation (–
.20). These negative relationships did not reject the hypotheses relating to the
simplex-ordered structure of the correlation matrix, however, because the direct
298 / Chatzisarantis, Hagger, Biddle, Smith, and Wang
Table 2 Meta-Analytic Estimates of Correlation Coefficients Between
Dimensions of PLOC
Correlation Error
No. of Sample coefficients estimates
studies size SE ME SE SD
Amotivation / External regulation 10 4198 .41 .53 .06 .24
Amotivation / Introjection 12 4927 .10 .14 .03 .13
Amotivation / Identification 10 4124 –.11 –.15 .07 .29
Amotivation / IM to know 4 1914 –.11 –.14 .05 .10
Amotivation / IM to accomplish 4 1914 –.16 –.20 .06 .13
Amotivation / IM to experience
stimulation 8 3996 –.20 –.27 .06 .22
External regulation / Introjection 17 7085 .34 .45 .04 .19
External regulation /Identification 16 6011 .08 .10 .06 .34
External regulation / IM to know 4 1914 .13 .16 .07 .16
External regulation / IM to accomplish 5 2016 .09 .11 .09 .25
External regulation / IM stimulation 18 7745 –.09 –.12 .06 .32
Introjection / Identification 17 6397 .40 .55 .02 .11
Introjection / IM to know 4 1914 .15 .19 .05 .10
Introjection / IM to accomplish 4 1914 .22 .29 .05 .11
Introjection / IM to experience
stimulation 17 7156 .31 .39 .03 .15
Identification / IM to know 4 1914 .36 .48 .07 .17
Identification / IM to accomplish 5 2016 .32 .43 .06 .16
Identification / IM to experience
stimulation 16 6011 .58 .75 .05 .25
IM to know / IM to accomplish 4 1914 .62 .75 .04 .10
IM to know / IM stimulation 5 2015 .55 .69 .04 .10
IM to accomplish / IM stimulation 5 2016 .62 .79 .08 .21
Note: IM = intrinsic motivation. Correlation coefficient corrected for: sampling error (SE);
Sampling error and measurement error (ME). Error estimates are standard errors (SE) and
standard deviations (SD). Underlined correlation coefficients are not significantly different
from zero at .001 alpha level.
effects of amotivation to identification, and of external regulation to intrinsic mo-
tivation, were smaller than the indirect effects of external regulation to identifica-
tion. These results therefore support the conclusion that the structure of correlation
matrices involving external regulation, introjection, and identification are indeed
simplex-ordered, and that this structure is independent from amotivation and in-
trinsic motivation, as would be predicted by proponents of self-determination theory
(Deci & Ryan, 1985; Ryan & Connell, 1989).
A Meta-Analysis of PLOC / 299
We examined the simplex-ordered structure of the correlation matrix,
operationalizing intrinsic motivation on three dimensions, by estimating the pa-
rameters of the model of Figure 1b. Unlike the model of Figure 1a, the fit of this
second model did not exceed recent criteria of good fit; but it did exceed older
criteria of good fit, given that the comparative fit index was .92 (Hu & Bentler,
1999). The Lagrange Multiplier Test indicated that the model’s fit would improve
significantly if three parameters were assumed to be greater than zero (chi-square
difference = 129.42, p < .001). These parameters were concerned with two paths
from amotivation to intrinsic motivation to accomplish, and to identification, as
well as with a path from external regulation to intrinsic motivation to experience
stimulation.
After these path coefficients were released, path-analytic estimates revealed
a positive indirect effect from external regulation to identification (indirect effect
= .30). In addition, the paths indicated by the Lagrange Multiplier Test were all
negative and their absolute values were smaller than the indirect effect of external
regulation on identification. Collectively, these results support the simplex-ordered
structure of correlation matrices including external regulation, introjection, and
identification. In addition, the strength of the structure is independent of the three-
dimensional concept of intrinsic motivation.
Mediating Effects of PLOC
Table 3 presents correlation coefficients, corrected for sampling and mea-
surement error, between PLOC, perceived competence, and intentions, as well as
standard errors and deviations of each pairwise correlation coefficient. The rela-
tively low standard errors suggested that, in spite of the small number of studies,
correlations were accurate and significantly different from zero at p < .001. An
exception was the correlation coefficient between external regulation and compe-
tence, which was statistically insignificant at the .001 alpha level. With the excep-
tion of standard deviations for correlation coefficients between competence and
intentions, and between amotivation and competence, all other standard devia-
tions were large.
To examine the mediating effects of PLOC on the relationship between per-
ceived competence and intentions, we conducted a path analysis using correla-
tions, corrected for sampling and measurement error, as an input matrix (Tables 2
and 3). In keeping with the path model of Figure 1, the model of Figure 2 is scale-
invariant because none of the parameters were fixed at a non-zero value (Cudeck,
1989). The only parameters that were fixed at a zero value were direct effects of
perceived competence on intentions. Finally, residual correlation coefficients be-
tween adjacent dimensions of PLOC were estimated because the important ante-
cedent variables of perceived autonomy and relatedness were missing (i.e., the
model is incomplete). Therefore the only hypothesis that was tested concerned the
mediating effects of PLOC in the perceived competence/intention relationship.
Note here that the power of the path model to detect a direct association between
competence and intentions is very low, given that the Type II error rates of the
correlation between competence and intentions are estimated to be .17 at the .05
alpha level (Field, 2001).
The comparative fit index for the model of Figure 2 was .97, and thus ex-
ceeded recent criteria for good fit (Hu & Bentler, 1999). Furthermore, the Lagrange
300 / Chatzisarantis, Hagger, Biddle, Smith, and Wang
Table 3 Meta-Analytic Estimates of Correlation Coefficients Between PLOC,
Perceived Competence, and Intentions
Correlation Error
No. of Sample coefficients estimates
studies size SE ME SE SD
Amotivation / Competence 5 2161 –.27 –.36 .07 .03
External regulation / Competence 7 3784 –.01 –.02 .07 .26
Introjection / Competence 8 4112 .24 .34 .04 .12
Identification / Competence 7 3024 .35 .48 .05 .18
IM to experience stimulation /
Competence 7 3784 .48 .62 .06 .20
Amotivation / Intentions 3 724 –.36 –.48 .07 .15
External regulation / Intentions 6 2454 –.15 –.20 .04 .13
Introjection / Intentions 7 2782 .23 .30 .04 .11
Identification / Intentions 6 1695 .39 .50 .06 .17
IM to experience stimulation /
Intentions 6 2454 .53 .67 .07 .19
Competence / Intentions 5 1332 .38 .47 .02 .00
Note: Correlation coefficient corrected for: sampling error (SE); Sampling error and measure-
ment error (ME). Error estimates are standard errors (SE) and standard deviations (SD).
Underlined correlation coefficients not significantly different from zero at .001 alpha level.
Multiplier Test did not indicate improvements in model fit if the direct effect of
perceived competence to intentions was estimated. With regard to parameters of
the model, there were statistically significant indirect effects of perceived compe-
tence to intentions via amotivation, intrinsic motivation to experience stimulation,
and introjection. The remaining indirect paths were not significant. The total effect
of perceived competence on intentions was .51. Therefore, results of the path analy-
sis support the existence of indirect effects of perceived competence on intentions
via amotivation, introjection, and intrinsic motivation to experience stimulation,
but not via external regulation and identification.
Discussion
The Self-Determination Continuum
The meta-analysis of correlations between dimensions of PLOC support two
conclusions pertinent to the self-determination continuum in sport and exercise.
The models presented in Figure 1a and 1b support the conclusion that the pattern
of correlations between external regulation, introjection, and identification con-
form to a simplex-ordered structure. However, a correlation matrix including mea-
A Meta-Analysis of PLOC / 301
sures of intrinsic motivation (unidimensional and three-dimensional intrinsic mo-
tivation) and amotivation is not, strictly speaking, a simplex-ordered matrix be-
cause the Lagrange Multiplier Test points out linear dependencies between
nonadjacent dimensions. Furthermore, estimation of nonhypothesized paths of
amotivation and intrinsic motivation, suggested by the Lagrange Multiplier Test,
did not attenuate the hypothesized indirect effects of external regulation to identi-
fication via introjection. These results therefore support the assumptions of self-
determination theory concerning the independence of internalization and
intrinsically motivated processes.
More specifically, it has been suggested that intrinsic motivation, amotivation,
and internalization are different processes (Deci et al., 1999a; Ryan & Deci, 2000).
Intrinsic motivation occurs when interest is the main factor that motivates behav-
ior. In contrast, internalization explains how individuals become motivated to en-
gage in an initially boring activity (Deci et al., 1994). During internalization,
experiences of enjoyment and interest are still important (Deci & Ryan, 1990);
however, internalization is expected to be phenomenologically aversive, at least
during the initial stages of interaction with the behavior. In the present meta-analy-
sis, the unpleasant nature of internalization processes is supported by the mediat-
ing effects of introjection on the relationship between external regulation and
identification that were not attenuated by amotivation and measures of intrinsic
motivation.
A second conclusion regarding the validity of the self-determination con-
tinuum is concerned with variability of study results. Results of primary studies
appear to be variable because none of the standard deviations of correlations be-
tween dimensions of PLOC were small (Standage, Duda, & Ntoumanis, 2003).
Variability of results from primary studies cannot be attributed to context or type
of PLOC measure, given that moderator analysis did not reveal any significant
results. However, the Type II error rates of this moderator analysis should be in-
flated because of the relatively small number of studies. Ideally, a moderator meta-
analysis of a larger set of studies will permit a statistically powerful moderator
analysis, using context (physical education, sport, physical activity), gender, and/
or age.
An important variable that might have contributed to variable results across
studies is concerned with study design. Few of the studies using PLOC were ex-
perimental (Standage, Treasure, Duda, & Prusak, 2003). Most of the designs were
cross-sectional or prospective and used self-report measures. Moreover, prospec-
tive studies were aimed at the prediction of repeated measures of behavior and did
not assess dimensions of PLOC and intentions repeatedly. From the perspective of
self-determination theory, the continuum is assumed to be a “developmental con-
tinuum,” meaning that with time comes a progressive change of various internal-
ization processes (Ryan & Deci., 2000). According to Li (1999), a developmental
continuum can be confirmed through panel designs, and by assessing individuals
over time while assessing PLOC at least twice (Hertzog & Nesselroade, 1987).
Such designs provide more accurate tests of a developmental self-determination
continuum and, consequently, may help deduce accurate inferences about the self-
determination continuum. For this reason, the present conclusions about the pro-
cess of internalization should be treated as tentative because of the cross-sectional
design of the studies.
302 / Chatzisarantis, Hagger, Biddle, Smith, and Wang
On the Antecedents and Consequences of PLOC
In addition to examining the structure of correlations between dimensions of
PLOC, the meta-analysis was used to examine associations of PLOC with ante-
cedent and outcome variables. Similar to correlations between dimensions of PLOC,
the meta-analysis of correlations between PLOC with perceived competence and
intentions revealed accurate but inconsistent results across studies. Such inconsis-
tency alerts researchers on the existence of possible moderator variables. Accu-
racy of correlations, however, supports the power of the meta-analysis to detect
statistically significant relationships.
Path analysis, using the correlation matrix corrected for sampling and mea-
surement error (see Table 3) as input matrix, supported the mediating effects of
PLOC on the relationships between perceived competence and intentions. Mediat-
ing effects of PLOC provide strong support of Vallerand’s (1997) hierarchical model
of intrinsic and extrinsic motivation. More specifically, Vallerand found that at a
contextual level of analysis such as a school setting, perceived self-determination
mediated effects of sport competence on intentions. However, that study used a
composite score, collapsing across dimensions of PLOC.
The present meta-analysis replicates the findings reported by Vallerand, in
addition to having used all the dimensions of PLOC. This analysis allows us to
conclude that the effects of perceived competence on intentions vary as a function
of PLOC. According to Deci and Ryan (1985), effects of perceived competence on
intrinsic motivation (and hence on intentions) can be expected when a controlling
aspect or informational aspect of internal events (such as competence) is salient to
the individuals. The present study confirms this proposition in the context of sport,
education, and leisure physical activity by showing that perceived competence
influences intentions via intrinsic motivation to experience stimulation, introjec-
tion, and amotivation (Ntoumanis, 2001a; Ryan, Koestner, & Deci, 1992).
Although relationships between perceived competence and dimensions of
PLOC are in line with theoretical predictions of self-determination theory, the re-
lationship between introjection and intentions is less well explained. In general,
most of the studies used Ajzen and Fishbein’s (1980) theoretical and operational
definition of intentions, which assumes that intentions represent volitional deci-
sions. However, the unique effect of introjection on intentions indicates that the
latter may not always represent volitional decisions, given that introjection repre-
sents behavior that is endorsed with a sense of compulsion. Therefore a final, but
tentative, conclusion that can be drawn from the results of the present meta-analy-
sis is that intentions confound volitional and forced decisions.
As a result, it can be argued that theories which represent volitional pro-
cesses through unidimensional measures of intentions contain an element of inde-
terminacy because unidimensional intentions confound volitional and forced
decisions. For this reason it may be appropriate to use alternative measures of
intentions, such as integrated intentions (Chatzisarantis, Biddle, & Meek, 1997),
which differentiate between volitional and forced decisions. Furthermore, it can
be suggested that measures of integrated intentions and PLOC may differ. Inte-
grated intentions indicate the degree to which a decision to engage in prospective
action is volitional. In contrast, the PLOC leaves time unspecified as it measures
one’s usual motivation in leisure time.
A Meta-Analysis of PLOC / 303
Conclusions
Results of the present meta-analysis support the existence of a self-determi-
nation continuum from external regulation to identification via introjection. Also,
the self-determination continuum appears to be independent from amotivation and
intrinsic motivation. Thus it can be concluded that internalization, intrinsic moti-
vation, and amotivation constitute qualitatively distinct processes. Moreover, re-
sults from the path analysis show that introjection and intrinsic motivation mediate
the effects of perceived competence on physical activity. Hence physical compe-
tence is a necessary but not sufficient condition for developing strong intentions
and an internal PLOC, given that perceived competence can induce introjections.
As for the apparent inconsistency of results, relationships between dimen-
sions of PLOC and between PLOC and perceived competence and intentions vary
greatly, even after accounting for differences in sample size and reliability be-
tween studies. Variation in study results is an indicator of possible moderator vari-
ables that the present meta-analysis might not have been powerful enough to detect,
due to the small number of studies available. However, the power of the present
meta-analysis is sufficient to detect statistical significance of hypothesized rela-
tionships that are greater than .30 at the .05 alpha level. Hence the failure to ex-
plain variability in study results is due to small differences that may represent the
influence of context on relationships between dimensions of PLOC. Therefore,
given the high power and low Type I error rates, which have been controlled by
reducing the alpha level to .001, the results of the present meta-analysis allow
confidence in the aforementioned conclusions.
References
*Denotes study used in the meta-analysis.
Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behavior.
Englewood Cliffs, NJ: Prentice Hall.
Bagozzi, R.P., & Kimmel, S.K. (1995). A comparison of leading theories for the prediction
of goal directed behaviours. British Journal of Social Psychology, 34, 437-461.
*Biddle, S., Soos, I., & Chatzisarantis, N. (1999). Predicting activity intentions using goal
perspective theories and self-determination theory approaches. European Psycholo-
gist, 4, 83-89.
*Briere, N., Vallerand, R., 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.
*Chantal, Y., Guay, F., & Martinova, T., & Vallerand, R.J. (1996). Motivation and elite
performance: An exploratory investigation with Bulgarian athletes. International Jour-
nal of Sport Psychology, 27, 173-182.
*Chatzisarantis, N.L.D., & Biddle, S.J.H. (1998). Functional significance of psychological
variables that are included in the Theory of Planned Behavior: A Self-Determination
Theory approach to the study of attitudes, subjective norms, perceptions of control
and intentions. European Journal of Social Psychology, 28, 303-322.
*Chatzisarantis, N.L.D., Biddle, S.J.H., & Meek, G.A. (1997). A Self-Determination Theory
approach to the study of intentions and the intention-behavior relationship in children’s
physical activity. British Journal of Health Psychology, 2, 343-360.
304 / Chatzisarantis, Hagger, Biddle, Smith, and Wang
*Chatzisarantis, N.L.D., Hagger, M.S., Biddle, S.J.H., & Karageorghis, C. (2002). The cog-
nitive processes by which perceived locus of causality predicts participation in physical
activity. Journal of Health Psychology, 7, 685-699.
Chaiken, S. (1980). Heuristic versus systematic information processing and the use of source
versus message cues in persuasion. Journal of Personality and Social Psychology,
39, 752-766.
Cudeck, C. (1989). Analysis of correlation matrices using covariance structure models.
Psychological Bulletin, 105, 317-327.
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., Koestner, R., & Ryan, R. (1999a). A meta-analytic review of experiments examin-
ing the effects of extrinsic rewards on intrinsic motivation. Psychological Bulletin,
125, 627-668.
Deci, E.L., Koestner, R., & Ryan, R.M. (1999b). The undermining effect is reality after all.
Extrinsic rewards, task interest and self-determination: Reply to Eisenberg, Pierce,
and Cameron (1999) and Lepper, Henderlong, and Gingras (1999). Psychological
Bulletin, 125, 692-700.
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. (1990). A motivational approach to self: Integration in personal-
ity. In R.D. Dienstbier (Ed.), Nebraska symposium on motivation: Perspectives on
motivation (Vol. 38, pp. 237-288). Lincoln: University of Nebraska Press.
Eisenberg, R., Pierce, D., & Cameron, J. (2000). Effects of reward on intrinsic motivation –
Negative, neutral, and positive: Comment on Deci, Koestner, and Ryan (1999). Psy-
chological Bulletin, 125, 677-691.
Field, A. (2001). A meta-analysis of correlation coefficients: A Monte Carlo comparison of
fixed and random effects methods. Psychological Methods, 6, 161-180.
*Georgiadis, M.M., Biddle, S.J.H., & Chatzisarantis, N.L.D. (2001). The mediating role of
self determination in the relationship between goal orientations and physical self
worth in Greek exercisers. European Journal of Sport Sciences, 1, 1-9
*Goudas, M., Biddle, S.J.H., & Fox, K. (1994). Perceived locus of causality, goal orienta-
tions and perceived competence in school physical education classes. British Jour-
nal of Educational Psychology, 64, 453-463.
*Goudas, M., Biddle, S.J.H., & Underwood, M. (1995). A prospective study of the relation-
ships between motivational orientations and perceived competence with intrinsic
motivation and achievement in a teacher education course. Educational Psychology,
5, 89-96.
*Hagger, M.S., Chatzisarantis, N.L.D., & Biddle, S.J.H. (2002). The influence of autono-
mous and controlling motives on physical activity intentions within the theory of
planned behaviour. British Journal of Health Psychology, 7, 283-297.
*Hamer, M., Karageorghis, C.I., & Vlachopoulos, S.P. (2002). Motives for exercise partici-
pation as predictors of exercise dependence among endurance athletes. Journal of
Sports Medicine and Physical Fitness, 42, 233-238.
Hertzog, C., & Nesselroade, J.R. (1987). Beyond autoregressive models: Some implica-
tions of the trait-state distinction for the structural modeling of developmental change.
Child Development, 58, 93-109.
A Meta-Analysis of PLOC / 305
Hu, L., & Bentler, M. (1999). Cut-off criteria for fit indexes in covariance structure analy-
sis: Conventional criteria versus new alternatives. Structural Equation Modeling: A
Multidisciplinary Journal, 6, 1-55.
Hunter, J., & Schmidt, F. (1990). Methods of meta-analysis: Correcting error and bias in
research findings. Newbury Park, CA: Sage.
Hunter, J., Schmidt, F., & Jackson, G. (1982). Meta-analysis: Cumulating research findings
across studies. Newbury Park, CA: Sage.
Joreskog, K. (1970). Analyzing psychological data by structural analysis of covariance
matrices. In D.H. Krantz, R.C. Atkinson, L.D. Luce, & D. Suppes (Eds.), Contempo-
rary developments in mathematical psychology (Vol. 2, pp. 1-56). San Francisco:
Freeman.
*Li, F. (1999). The Exercise Motivation Scale: Its multifaceted structure and construct
validity. Journal of Applied Sport Psychology, 11, 97-115.
*Li, F., & Harmer, P. (1996). Testing the simplex assumption underlying the sport motiva-
tion scale: A structural equation analysis. Research Quarterly for Exercise and Sport,
67, 396-405.
Markland, D., & Hardy, L. (1997). On the factorial and construct validity of the intrinsic
motivation inventory. Research Quarterly for Exercise and Sport, 68, 20-32.
*Mullen, E., Markland, D., & Ingledew, D.K. (1997). A graded conceptualization of self-
determination in the regulation of exercise behavior: Development of a measure us-
ing confirmatory factor analysis. Personality and Individual Differences, 23, 745-752.
Nicholls, J.G. (1984). Conceptions of ability and achievement motivation. In R. Ames & C.
Ames (Eds.), Research on motivation in education (Vol. 1, pp. 39-73). New York:
Academic Press.
*Ntoumanis, N. (2001a). A self-determination approach to the understanding of motivation
on physical education. British Journal of Educational Psychology, 71, 225-242.
*Ntoumanis, N. (2001b). Empirical links between achievement goal theory and self-deter-
mination theory in sport. Journal of Sports Sciences, 19, 397-409.
*Pelletier, L., Fortier, M., Vallerand, R., Tuson, K., & Blais, M. (1995). Toward a new
measure of intrinsic motivation, extrinsic motivation, and amotivation in sports: The
Sport Motivation Scale (SMS). Journal of Sport & Exercise Psychology, 17, 35-53.
Ryan, R.M., & Connell, J.P. (1989). Perceived locus of causality and internalization. Jour-
nal of Personality and Social Psychology, 5, 749-761.
Ryan, R.M., & Deci, E.L. (2000). Self-determination theory and the facilitation of intrinsic
motivation, social development, and well being. American Psychologist, 55, 54-56.
Ryan, R.M., Koestner, R., & Deci, E.L., (1992). Ego-involved persistence: When free-choice
behaviour is not intrinsically motivated. Motivation and Emotion, 15, 185-205.
Schmidt, F. (1996). Statistical significance testing and cumulative knowledge in psychol-
ogy: Implications for training researchers. Psychological Methods, 2, 115-129.
Schmidt, F., & Hunter, J. (1999). Comparison of three meta-analysis methods revisited: An
analysis of Johnson, Mullen and Sallas (1995). Journal of Applied Psychology, 84,
144-148.
*Standage, M., Duda, J.L., & Ntoumanis, N. (2003). A model of contextual motivation in
physical education: Using constructs of self determination and achievement goal
theories to predict physical activity intentions. Journal of Educational Psychology,
95, 97-110.
306 / Chatzisarantis, Hagger, Biddle, Smith, and Wang
*Standage, M., Treasure, D.C., Duda, J.L., & Prusak, K. (2003). Validity, reliability, and
invariance of the Situational Motivation Scale (SIMS) across diverse physical activ-
ity contexts. Journal of Sport & Exercise Psychology, 25, 19-43.
Tett, R., & Meyer, J. (1993). Job satisfaction, organizational commitment, turnover inten-
tion, and turnover: Path analyses based on meta-analytic findings. Personnel Psy-
chology, 2, 259-293.
Vallerand, R.J. (1997). Toward a hierarchical model of intrinsic and extrinsic motivation. In
M. Zanna, Advances in experimental social psychology (pp. 271-360). New York:
Academic Press.
*Vallerand, R.J., & Losier, G. (1994). Self-determined motivation and sportsmanship ori-
entations: An assessment of their temporal relationship. Journal of Sport & Exercise
Psychology, 16, 229-245.
Vallerand, R.J., Pelletier, L.G., Blais, M.R., Briere, N.M., Senecal, C., & Vallieres, E.F. (1992).
The academic motivation scale: A measure of intrinsic, extrinsic, and amotivation in
education. Educational and Psychological Measurement, 52, 1003-1017.
*Wang C.K.J., Chatzisarantis, N.L.D., Spray, C.M., & Biddle, S.J.H. (2002). Achievement
goal profiles in school physical education: Differences in self determination, sport
ability beliefs, and physical activity. British Journal of Educational Psychology, 72,
433-445.
Whitener, E. (1992). Confusion of confidence intervals and credibility intervals in meta-
analysis. Journal of Applied Psychology, 75, 315-321.
Manuscript submitted: October 9, 2001
Revision accepted: March 25, 2003
