A test of the trans-contextual model of motivation in Greek high school pupils.

Barkoukis, Vassilis ; Hagger, Martin S.

Research evidence has documented the positive effect of regular vigorous physical activity on the physical (Gilliam & MacConnie, 1985) and mental health (Callas & Taylor, 1994) of young people. However, there is also considerable body of research highlighting that young people in industrialized societies do not engage in sufficient physical activity thought to promote health (Centers for Disease Control and Prevention, 2003). This has led to the publication of a number of recommendations regarding desirable levels of physical activity children and adolescents (Sallis & Patrick, 1994). Health professionals involved in developing campaigns and curricula to promote physical activity in young people have sought to develop effective interventions to increase physical activity levels in young people. The use of existing networks such as school physical education (PE) is considered crucial in this regard (Fox & Biddle, 1989). PE is an ideal context in which children and adolescents can learn more about healthy lifestyles, particularly physical activity. Campaigns promoting physical activity need to be oriented about theoretical research to identify the salient predictors of physical activity participation in young people (Biddle, Gorely, & Stensel, 2004). Current research is focused on the theoretical approaches that describe the mechanisms that influence motivation for participation in physical activity (Hagger & Chatzisarantis, in press; Vallerand, 2007).

An important question raised from research concerns whether factors affecting motivation in PE could increase motives for participation in physical activities during leisure-time (Ntoumanis, 2005; Standage, Duda, & Ntoumanis, 2003). The present study aims to address this question by adopting an integrated theory of motivation, known as the trans-contextual model (Hagger, Chatzisarantis, Culverhouse, & Biddle, 2003). According to this model, school pupils' perceptions that their PE teacher supports their autonomy during lessons are positively related to their motivation in a PE context but also motivation to engage in physical activities in their leisure-time outside of school. In order to understand this complex motivational sequence, the trans-contextual model integrates hypotheses from three theories of motivation; self-determination theory (Deci & Ryan, 1985), the hierarchical model of intrinsic motivation (Vallerand, 2007), and the theory of planned behavior (Ajzen, 1985). The subsequent sections will outline the major theories and then illustrate how they can provide complimentary explanations of the proposed motivational sequence.

Self-determination Theory

The central tenet of self-determination theory is the distinction between self-determined or autonomous and non-serf-determined or controlled forms of motivation (Deci & Ryan, 1985, 2000). The experience of autonomous is characterized by a sense of personal origin, self-regulation, and a sense that one's actions are fully endorsed by the self. Important outcomes of autonomous functioning are perceptions of interest, enjoyment, satisfaction and a sense of choice. Behaviors that people engage in for autonomous motives are those that satisfy the basic needs for autonomy, competence and relatedness. On the other hand, people that act for controlled reasons engage in behavior but feel tension and conflict with their behavior and perceptions. This typically arises because their actions are contingent on external reinforcers such as obtaining rewards or avoiding punishment or internal conflicts, such as perceptions of obligation or guilt. Importantly, autonomous forms of motivation are associated with behavioral persistence and adaptive outcomes while controlled motives are akin to desistance, avoidance, and maladaptive outcomes. Several reviews and recta-analyses have illustrated that autonomous forms of motivation are associated with increased effort, persistence, and satisfaction among individuals engaging in tasks in physical activity contexts (Cbatzisarantis, Hagger, Biddle, Smith, & Wang, 2003; Hagger & Chatzisarantis, 2008).

Hierarchical Model of Intrinsic Motivation

Vallerand (2007) adopted several aspects of the self-determination theory and developed a hierarchical model of intrinsic motivation. This model focuses on the role of environment and motivation at different levels of generality to describe the consequences or outcomes of motivated behavior (Vallerand, 2007). According to the model, global motivation lies at the apex of the hierarchy. Global motives are general, stable, represent motivational dispositions, and reflect the basic needs for self-determination or autonomy. Global motives are thought to influence motivation in many different contexts in a top-down fashion (Guay, Mageau, & Vallerand, 2003).

Motivation at the contextual level is located at an intermediate level in the hierarchy, subordinate to global motivation, and represents motivation in specific contexts, such as the academic, social, and physical contexts. These motives are less stable and can be influenced by external factors (social and environmental) as top-down effects from global level motives. An important prediction of the model for the contextual level is the interaction between different contexts (Vallerand, 2007). For example, intrinsic motives in an academic context (i.e., motivation in PE) could be transferred as intrinsic motives in a social or sport context (i.e., leisure motivation).

Situational motivation lies below contextual motivation in the hierarchy (Vallerand, 2007). Situational motivation refers to the type of motivation operating in specific situations. For example, the motivational states operating in the physical context might include motivation to participate in activities in a PE lesson or training at a sports club. The type of motivation experienced by individuals in each of these situations is partly determined by their contextual motivational orientation as well as the environmental conditions affecting motivation in that situation. The model presumes that contextual motives mediate the top-down influence of global-level motives on situational motives and vice versa (Guay et al., 2003).

The Importance of Autonomy Support

According to self-determination theory and the hierarchical model, an environment that fosters autonomous motivation in an educational context will lead to persistence with educational tasks and behaviors in the absence of external reinforcement (Deci & Ryan, 1985, 2000; Reeve, 2002). The mechanism behind the positive influence of autonomy support on behavior is ascribed to the positive influence that autonomy support has on autonomous forms of motivation (Reeve, 2002). Autonomous motives serve to mediate the effect of autonomy support on behaviors and are therefore an important strategy to develop autonomous reasons for acting (Hagger et al., 2003).

The Theory of Planned Behavior

The theory of planned behavior is a social-cognitive theory thought to adequately explain intentional behavior (Ajzen, 1985). According to the theory, intention to perform a behavior at a specific time point is the strongest predictor of this behavior. Intention represents the plans of the individual to engage in a target behavior in a specific context and time. Intention is influenced by three social-cognitive variables. These variables are attitudes toward behavior, reflecting the individual's positive and negative personal beliefs about the target behavior; subjective norm; representing the social pressure that significant others present to the actor regarding his/her performance of the target behavior; and perceived behavioral control, which refers to the individual's perception about his/her ability to perform the behavior.

The basic hypotheses of the theory have been supported in a large number of behavioral domains (Armitage & Conner, 2001), including exercise and physical activity (Hagger, Chatzisarantis, & Biddle, 2002b). However, criticisms have been raised about the theory as it does not account for all of the variance in intention and behavior (Conner & Armitage, 1998), and has limited predictive utility for long-term behaviors due to the boundary condition of stability (Chatzisarantis & Biddle, 1998). The boundary condition stipulates that the prediction of behavior is limited due to the inherent instability of the situational constructs contained in the model. Such constructs are open to change from the advent of additional information, compromising the predictive utility of the model (Sheeran, Orbell, & Trafimow, 1999).

Researchers have suggested that this theory should be augmented to incorporate more generalized constructs from self-determination theory and the hierarchical model to explain the origins of its situation-specific constructs (Chatzisarantis & Biddle, 1998; Deci & Ryan, 1985; Hagger et al., 2003). Such theoretical modifications have been justified by a number of authors who argue that the decision-making process mapped by the theory of planned behavior reflects situation-level motivation to engage in behaviors. Furthermore, the direct antecedents of intentions, namely attitudes and perceived behavioral control, reflect autonomous reasons for forming intentions or plans to act (Sheeran, Norman, & Orbell, 1999). The theory of planned behavior therefore, has potential to integrate with Vallerand's (2007) hierarchical model as its constructs, particularly attitudes, perceived behavioral control, and intentions, reflect motivational orientations at the situational level.

Integration of the Theories: The Trans-Contextual Model

Self-determination theory and the theory of planned behavior may offer complementary explanations of motivated behavior. Self-determination theory provides explanations of the environmental and interpersonal states that give rise to behavioral tendencies in given contexts, Vallerand's (2007) hierarchical model suggests that context level motivation influences motivation at the situational level, and the theory of planned behavior represents the autonomous decision-making process towards behavioral engagement at the situational level. Recent evidence has supported these notions in studies of health behavior including physical activity. For example, autonomous motives affect attitudes, and attitudes, in turn, influence autonomous motives on intention to perform a behavior (Hagger & Armitage, 2004; Hagger, Chatzisarantis, & Biddle, 2002a). Hugger and colleagues (2002a) employed this theoretical synthesis and provided empirical support for the effect of autonomous motives on the theory of planned behavior constructs. They found that attitudes and perceived behavioral control mediated the effect of autonomous motives on behavioral intention and physical activity behavior in accordance with Ajzen's (1985) original conceptualization of the theory of planned behavior. There have since been a number of studies adopting this integrated approach and a meta-analysis of 12 studies has supported the integration of the theory of planned behavior and self-determination theory (Hagger & Chatzisarantis, 2007).

Extending these findings, Hagger et al. (2003) examined whether the pattern of influence identified previously operated across contexts. Specifically, the hypothesis that there could be interplay between environmental contingencies and motivation across contexts as well top-down from a contextual to a situational level was applied to the question of whether motivation can be transferred across contexts. The resulting model, known as the trans-contextual model, explicitly illustrates how an environment perceived to be autonomy supportive can promote autonomous motivation in a different context and whether that motivation can be converted into intentions to act and actual behavior. The model was tested in the PE and leisure-time contexts with physical activity as the target behavior. According to the model, autonomy support in PE lessons would have positive influence on the autonomous motives in the same context (Ntoumanis, 2005; Standage et al., 2003; Wilson & Rodgers, 2004). Furthermore, autonomy support in PE lessons could affect autonomous motives in a non-educational setting-leisure-time physical activity. This transference of motivational orientations and perceived autonomy support is explicitly stated to occur across contexts in Vallerand's (2007) hierarchical model. There is interplay between motivational orientations across life contexts at the contextual level as individuals use information regarding their level of autonomy towards behaviors in related contexts as a source of information when forming motivational orientations in subsequent contexts.

In addition, as the top-down processes of the hierarchical model imply, autonomous motives toward leisure-time physical activity were hypothesized to affect motivation to act in specific situations. This top-down process was represented by the effects of autonomous motives on the situational variables from the theory of planned behavior, particularly intentions. Furthermore, these motives are expected to mediate the influence of perceived autonomy support on intention to perform a behavior, as well as the behavior itself (Vallerand, 2007). The preliminary investigation of the model supported the proposed motivational sequence with no direct effects of perceived autonomy support or autonomous motives on intention and behavior (Hagger et al., 2003). Since then, a number of studies have supported the premises of the trans-contextual model and a recent meta-analysis of eight tests of the pattern of influence among the study constructs provided support for the proposed sequence spanning from autonomy support in a PE setting to regular physical activity in a leisure-time setting (Hagger & Chatzisarantis, 2007). The model provides a comprehensive framework to explain how autonomy support in an educational context is translated into self-determined behavior in another context through its influence on autonomous motives from self-determination theory and constructs from the theory of planned behavior.

The Present Study

The present study aims to examine the motivational sequence proposed by the transcontextual model in a sample of Greek high-school pupils. Given that the trans-contextual model is a recently-developed model of motivation, it is not known whether the proposed effects in the model are unique to the culture in which it was developed or are applicable across populations with different cultural backgrounds. There is evidence to suggest that the processes highlighted by models of motivation like the trans-contextual model may be universal and, therefore, applicable to all adolescents in schools regardless of the cultural values they endorse. Indeed, cross-cultural research adopting self-determination theory and the theory of planned behavior have supported the generalizability of the model structure and the pattern of relationships in cultures that endorse different values. For example, Chirkov and Ryan (2001) found that perceived autonomy support had a significant influence on motivation and academic well-being in Russian students, who tend to adopt a more collectivist cultural orientation. In addition, Bagozzi, Lee, and Van Loo (2001) also found that the factor structure and pattern of influence of the theory of planned behavior was invariant across cultures that have predominantly individualist and collectivist orientations. Given this evidence, and the premise put forward by theorists that the processes mapped in motivational models are universal, the present study aims to examine the trans-contextual model in the Greek culture. This is a national group that endorses more collectivist cultural values relative to the nations in which the model was originally developed (United Kingdom) that typically endorses individualist norms (Oyserman, Coon, & Kemmelmeier, 2002). It is hypothesized that the major predictions of the trans-contextual model outlined in the previous section will be supported in this sample and will extend the literature in this area by providing some initial support for the generalizability of these processes across cultures.

Method

Sample

The sample of the study consisted of 183 adolescents (91 girls, 92 boys; M age = 13.93, SD = 0.78) from a co-educational high school in an urban city of Northern Greece. According to the school director the socioeconomic status of the adolescents could be described as middle class. The majority of the participants were native Greek nationals, while a small minority were immigrants from countries of the former Soviet Union. These adolescents were second-generation immigrants with Greek as their first language. The school director was informed on the purpose of the study and gave permission for the data collection. Consent from the participants' parents was also obtained via a letter sent home with the participants outlining the study prior to data collection. A pre-printed form was provided for parents to sign and return to the class teacher if they did not want their child to participate in the study. No forms were returned in any case. Children were also informed that participation in the study was voluntary and given the opportunity not to complete the questionnaires at each point of data collection without prejudice. All children opted to complete the questionnaire.

Measures

Perceived autonomy support in PE. Autonomy support in PE was measured using an initial version of the Perceived Autonomy Support Scale for Exercise Settings (PASSES; Hagger et al., 2007). The measure asks respondents to rate the extent that significant authoritative others support their autonomy in a given context. For the present study the authoritative others were specified as the participants' PE teachers. The scale has 15 items (e.g., "I feel that my PE teacher makes sure I really understand the goals of the lesson and what I need to do"). Responses were given on seven-point Likert scales ranging from 1 ("strongly disagree") to 7 ("strongly agree").

Autonomous motivation in PE and leisure-time. A modified version of Ryan and Connell's (1989) perceived locus of causality (PLOC) scale was used to measure autonomous motivation in PE. The scale measures four types of motivation (1), each varying in the degree of autonomy on a continuum ranging from high to low autonomy. Participants were presented with a common stem: "Why do you participate in PE?", followed by eight reasons, two from each regulation style from the perceived locus of causality: intrinsic motivation (e.g., "because PE is fun"), identified regulation (e.g., "because I value PE"), introjected regulation (e.g., "because I will feel ashamed if I do not"), and extrinsic regulation (e.g., "because important others want me to do it"). Responses were measured on four-point Likert-typo scales ranging from 1 ("not true at all") to 4 ("very true"). The construct and discriminant validity of the PLOC measure in a PE context has been reported in previous studies (Hagger et al., 2003).

Autonomous motivation in leisure-time physical activity was measured using the Behavioral Regulations in Exercise Questionnaire (BREQ, Mullan, Markland, & Ingledew, 1997). Participants were asked "Why do you participate in active sports and/or vigorous physical activities in your spare time?" followed by three reasons from each regulation style. The items measuring intrinsic motivation (e.g., "I exercise because it is fun"), identified regulation (e.g., "I exercise because it is important to make the effort"), introjected regulation (e.g., "I exercise because I will feel guilty if I do not"), and external regulation (e.g., "I exercise because others say I should") were assessed on seven-point Likert-type scales ranging from 1 ("not true at all") to 7 ("very true").

The theory of planned behavior questionnaire. Constructs from the theory of planned behavior questionnaire were measured using standardized items developed according to published guidelines (Ajzen, 2003). Three items measured behavioral intentions (e.g., "I intend to do active sports and/or vigorous physical activities in the next 5 weeks ...") on seven-point Likert scales anchored by 1 ("strongly agree") and 7 ("strongly disagree"). Attitudes were assessed in response to the following question: "Participating in active sports and/or vigorous physical activities during my leisure-time in the next five weeks is ..." Responses were measured on four seven-point semantic differential scales with the following bipolar adjectives: bad-good, harmful-beneficial, unenjoyable-enjoyable, and useful-useless. Subjective norms were measured by two items (e.g., "People important to me think that I should do active sports and/or vigorous physical activities during my leisure-time in the next 5 weeks") on seven-point scales with 1 ("strongly disagree") and 7 ("strongly agree") as endpoints. Perceived behavioral control was assessed through three items (e.g., "I feel in complete control over whether I do active sports and/or vigorous physical activities in my leisure-time in the next 5 weeks") measured on seven-point Likert scales ranging from 1 ("no control") to 7 ("complete control").

Self-reported physical activity behavior. Physical activity behavior was assessed by an adapted version of Godin and Shephard's (1985) Leisure-Time Exercise Questionnaire (LTEQ). Participants were initially presented with the definition of vigorous physical activity given by Godin and Shephard. Vigorous physical activity was defined as all active sports and physical activities, but not low-level activities or walking to school. Participants were requested to consider, on average, how many times per week they engaged in vigorous physical activity over the previous five weeks, recorded on a single seven-point scale. Independent evaluations of the LTEQ have found it to be valid, reliable, and easy to administer. It displays satisfactory concurrent validity with gold-standard measures of physical activity such as heart-rate monitoring (Jacobs, Ainsworth, Hartman, & Leon, 1993) and this version has been found effective in estimating physical activity levels in studies using social-cognitive models (Hagger et al., 2003; Hagger, Chatzisarantis, & Harris, 2006).

Cultural orientation. A sub-sample of the participants recruited at the first wave of data collection (N = 118) completed an abbreviated version of Triandis' (1995) individualism-collectivism scale (2). This nine-item measure was used to provide an estimate of the general cultural orientation of the sub-sample. The scale comprises four items measuring individualism (e.g., "I would rather make an important decision by myself than discuss it with my friends") and five items measuring collectivism (e.g., "I can count on my relatives for help if I find myself in any kind of trouble"). Responses were given on seven-point Likert scales with 1 ("not at all") and 7 ("very much") as endpoints.

Translation of the Measures

A back translation technique was used to translate the measures of the study into the Greek language (Brislin, 1986). In the first instance, the study measures were translated from English to Greek by a proficient translator. Two other proficient translators then translated the Greek version back into English. These versions were then compared with the original and the appropriate corrections were made to the Greek version. Next, this modified version was back-translated again into English and compared to the original version for further inconsistencies and biases. This procedure was repeated until there were no inconsistencies between the translated and original versions.

Research Design

A three-wave prospective design was implemented to allay common method variance and for the prospective prediction of behavior in accordance with the theory of planned behavior. At the first wave of data collection, the autonomy support in PE (Deci & Ryan, 2003) and autonomous motivation in PE (Ryan & Connell, 1989) measures were administered. One week later at a second wave of data collection, the measures of the theory of planned behavior components (Ajzen, 1985) and autonomous motivation during leisure-time (Mullah et al., 1997) were administered to the participants. The one-week delay ensured that the measures of autonomous motivation in the two contexts were not contaminated by common method variance associated with introducing measures with similar themes simultaneously. The third wave took place five weeks later and involved the administration of the self-report measure of physical activity behavior. A five-week period was used because it permits long-range prediction, a desirable outcome in motivational models (Hagger et al., 2003).

Researchers conducted the data collection in quiet classroom conditions. Participants were informed that they would be asked to complete a series of short questionnaires over the coming weeks as part of a survey on young people. Participants were separated so that they could not copy or discuss responses. All of the questionnaires were completed anonymously to preserve confidentiality. Measures across time-points were matched using birth date and gender.

Results

Participants

Attrition of participants across the three waves due to absences and incomplete questionnaires resulted in a final sample size of 93 high school pupils across the three waves (57 girls, 36 boys; Mage = 13.99, SD = 0.80). Scores on the Triandis (1995) abbreviated individualism-collectivism scale indicated that participants' average score on the collectivism scale was significantly higher than their score on the individualism scale (t(117) = 2.19, p < .05). This is consistent with cross-cultural data indicating that Greek nationals tend to favor a collectivist orientation relative to other Western European nations (Oyserman et al., 2002). This is in stark contrast to the predominantly individualist cultural orientation prevalent in the British culture where the trans-contextual model was developed (Hagger et al., 2003).

Preliminary Analysis

In order to maximize parsimony in the hypothesised model and to minimize the number of constructs in subsequent analyses, we calculated a relative autonomy index (RAI) from the PLOC components in both contexts, as recommended by other researchers (Guay et al., 2003). Weights were assigned to each participant's intrinsic motivation (+2), identified regulation (+1), introjected regulation (-1), and extrinsic regulation (-2) score, and the RAI was the composite of these weighted scores. The RAI therefore reflected participants' relative level of autonomous motivation in the PE and leisure-time physical activity contexts.

Descriptive statistics and zero-order correlations among the perceived autonomy support, RAI composites in both contexts, the theory of planned behavior variables, and physical activity behavior are given in Table 1. Correlations between perceived autonomy support and the RAI composites in both contexts were significant and positive. Perceived autonomy support was also significantly correlated with the theory of planned behavior constructs, with the exception of the subjective norms construct. The RAI for the leisure-time context was also significantly related to the theory of planned behavior variables and physical activity behavior, while the RAI for the PE context was not related to any of the theory of planned behavior constructs or physical activity behavior. Finally, measures displayed satisfactory levels of internal consistency, with the exception of the subjective norms construct. However, previous research has confirmed that measures of this construct generally do not achieve satisfactory reliability despite adhering to recommended guidelines for their development (Ajzen, 2003; Hagger, Chatzisarantis, & Biddle, 2001). Therefore, the results pertaining subjective norms should be interpreted with caution.

Main Analysis

Study hypotheses were examined by path analysis using manifest rather than latent variables and a simultaneous process with the EQS v. 6.0 computer program. Model fit was evaluated by means of the goodness-of-fit Chi-square, which should be non-significant for a well-fitting model, and other relative indexes of good fit, namely the comparative fit index (CFI), which should exceed .95, and the root mean square error of approximation (RMSEA) which should be less than .08 (Hu & Bentler, 1999). In the model, the hypothesized pattern of relationships among the trans-contextual model constructs proposed by Hagger et al. (2002a) was tested by including unidirectional free parameters (paths) mapping out the motivational sequence.

In the specified model, perceived autonomy support in PE was made to predict autonomous motives in a leisure-time physical activity context, as captured by the RAI, directly and indirectly via the mediation of autonomous motives in a PE context. In addition, the model proposed that autonomous motives in a PE context would predict attitudes and perceived behavioral control via the mediation of autonomous motives in a leisure-time context. In addition, paths from autonomous motives in a leisure time context to physical activity intentions via the mediation of attitudes and perceived behavioral control in a leisure-time context were included. There was also an indirect path from autonomous motives in a leisure-time context to physical activity intentions mediated by subjective norms. Finally, attitudes, subjective norms, and perceived behavioral control were made to predict physical activity behavior through the mediation of intentions.

Examination of the fit indexes revealed that the hypothesized model reproduced the observed covariance matrix satisfactorily ([chi square] = 6.71, df = 5; p = .24; CFI = .99; RMSEA = .06). The standardized path coefficients for the final model are given in Figure 1. Turning first to the effects of perceived autonomy support in PE in the model, as expected, perceived autonomy support predicted autonomous motives in a PE context ([beta] = .24, p < .01). In addition, there was also a direct effect of perceived autonomy support on autonomous motives in a leisure-time context ([beta] = .39, p < .01). Furthermore, the anticipated effect of autonomous motives in a PE context on autonomous motives was non-significant. So autonomous motives in PE did not mediate the effect of perceived autonomy support in a leisure-time context, as hypothesized. Furthermore, the non-significant direct path between the autonomous motivation constructs across contexts indicated that autonomous motivation in leisure time could not mediate the effect of autonomous motivation in a PE context on attitudes, subjective norms, and perceived behavioral control. There were no other direct effects of perceived autonomy support or autonomous motives in PE in the model.

Next we focus on the effects of autonomous motives in leisure-time on the theory of planned behavior constructs. Autonomous motives in leisure-time significantly influenced attitudes ([beta] = .54, p < .01), subjective norms ([beta] = .28, p < .01), and perceived behavioral control ([beta] = .53, p < .01) as hypothesized. There were no direct effects of autonomous motives on intention, which is congruent with previous tests of the trans-contextual model (Hagger et al., 2003) and the effects of self-determined motives on the theory of planned behavior constructs (Hagger et al., 2002a; Wilson & Rodgers, 2004). However, there was a significant and positive zero-order correlation between the autonomous motives construct and intentions (r = .45, p < .01; see Table 1). This suggests that the attitude, subjective norm, and perceived behavioral control constructs mediated the impact of autonomous motives on intentions.

In order to test this premise formally, we adopted the criteria of Baron and Kenny 0986). There are four criteria that require satisfaction in order for mediation to be supported: (i) the dependent variable (in this case, intention) should be correlated with the independent or predictor variable (autonomous motives in leisure-time); (ii) the mediator(s) (attitude/subjective norms/perceived behavioral control) should be correlated with the independent variable; (iii) the mediator(s) should have a significant unique effect on the dependent variable when it is included alongside the independent variable in a multivariate test of these relationships; and (iv) the effect of independent variable on the dependent should be significantly attenuated or nullified when the mediator(s) is/are included as an independent predictor(s) of the dependent variable (Baron & Kenny, 1986). In the present study, intention was significantly correlated with the autonomous motives construct in a leisure time context and the attitude, subjective norm, and perceived behavioral control constructs were also significantly correlated with autonomous motives, supporting the first two criteria for mediation. Further, attitude and perceived behavioral control had significant unique effects on intention, but subjective norms did not. Therefore subjective norms had to be rejected as a potential mediator, but the attitude and perceived behavioral control constructs satisfied the third criterion for mediation. Finally, removing the direct effect of attitude and perceived behavioral control on intention independently in the path model resulted in an increase in the direct effect of autonomous motives on intention in both cases, satisfying the final criterion for mediation. For completion, a Sobel (1982) test indicated that the indirect effect of autonomous motives on intention via attitudes and perceived behavioral control was significant ([beta] = .40, p < .01).

Turning next to the relationships between the theory of planned behavior variables, results showed that attitudes ([beta] = .40, p < .01) and perceived behavioral control ([beta] = .35, p < .01) were significant predictors of intention, as predicted. However, subjective norms did not significantly predict intentions, so this hypothesis had to be rejected. Further, intention significantly predicted leisure-time physical activity behavior. There were no direct effects of attitude or perceived behavioral control on behavior. It seemed that intentions mediated the effects of attitude and perceived behavioral control on behavior. A formal test of this was conducted using Baron and Kenny's (1986) criteria for mediation. From Table 1 it can be seen that attitude and perceived behavioral control are significantly correlated with leisure-time physical activity behavior and attitude and perceived behavioral control are significantly correlated with intention. These relationships confirmed the first two criteria for mediation. In order to confirm the final two criteria, we released free parameters in an alternative model between attitude and behavior and perceived behavioral control and behavior. Despite the inclusion of these paths, the intention construct remained the only significant predictor of behavior in the model, satisfying the third criterion for mediation. Next, fixing the intention-behavior relationship to be zero and re-estimating the model resulted in a restoration of the significant direct effects of attitude and perceived behavioral control on intention. Finally, Sobel tests confirmed the significant indirect effects of attitude ([beta] = .32, p < .01) and perceived behavioral control ([beta] = .28, p < .01) on behavior. Thus it can be concluded that, in accordance with Ajzen's (1985) hypothesis, intention mediated the influence of attitude and perceived behavioral control on leisure-time physical activity behavior.

Finally, it is important to examine the overall effects of the distal autonomy support and autonomous motivation constructs on intentions and behavior. Although there were no direct effects, there was a significant total effect of perceived autonomy support in PE on intentions ([beta] = .23, p < .01) and leisure-time physical activity behavior ([beta] = .23,p < .01) in the model. In addition, there was a significant total effect of autonomous motives in a leisure-time physical activity context on intentions ([beta] = .45, p < .01) and behavior ([beta] =. 19, p < .01). Overall, the model accounted for 22.21% of the variance in intentions and 44.76% of the variance in leisure-time physical activity behavior.

Discussion

The present study tested a trans-contextual model of motivation (Hagger et al., 2003; Hagger & Chatzisarantis, 2007) in Greek high school children. The model integrated three theories of motivation and a motivational sequence was proposed in which perceived autonomy support in a PE context would influence autonomous motivation, intentions, and behavior in a leisure-time context. The motivational sequence was generally supported in this sample of Greek children, with perceived autonomy support in PE related to autonomous motivation in a leisure time context. Autonomous motives in a leisure time context predicted intentions, but only via the mediation of attitudes and perceived behavioral control towards doing leisure-time physical activity. Intentions predicted behavior and mediated the influence of attitudes and perceived behavioral control in leisure-time physical activity behavior. Only two hypothesized effects based on theory and previous fundings with the trans-contextual model were not significant; the effect of autonomous motives in PE on autonomous motives in leisure-time and the effect of subjective norms on intention. The proposed motivational sequence was therefore supported to a great extent and the model explained a substantial proportion of the variance in leisure-time physical activity behavior.

Results from the present study support the use of the trans-contextual model in a different cultural group. One of the key premises of self-determination theory is that the autonomous drive to pursue behaviors that satisfy psychological needs are universal and salient to all cultures, regardless of their orientation on cultural values such as individualism and collectivism (Chirkov & Ryan, 2001). Results indicate that the trans-contextual model exhibits satisfactory goodness-of-fit with the present data from a sample of Greek students. The pattern of results in this test of the trans-contextual model in a cultural group that tends to endorse more collectivist cultural values, as illustrated by the abbreviated Triandis (1995) individualism-collectivism scale, compares favorably with that from a British culture (Hagger et al., 2003). These data suggest that the processes identified by Hagger et al. (2003) might also exhibit consistency across cultures and corroborated findings from the meta-analysis of the trans contextual model (Hagger & Chatzisarantis, 2007). Further cross-cultural replications and comparisons of the model are required to corroborate current findings (Hagger & Chatzisarantis, 2007).

[FIGURE 1 OMITTED]

An important funding of the present study was that perceived autonomy support in a PE context had a significant effect on leisure-time physical activity behavior, but only via a motivational sequence that included contextual-level motivation in a leisure-time context and the theory of planned behavior constructs. Therefore, adolescents in a PE context that perceive their teachers to provide opportunities and choice for their activities and help them to set personal goals are those more likely to report higher motivation to participate in physical activities outside of school. In turn, they are also more likely to form intentions to engage in future physical activity behavior, and engage in actual physical activity behavior.

This motivational sequence provides some salient information about the processes that underpin motivation and behavior from an educational context to an extra-curricular context. It supports Deci and Ryan's (1985, 2000) and Reeve's (2002) proposals that autonomy support is influential in fostering intrinsic motivation. However, this model is pioneering in that it supports this link across different contexts. The model also supports Hagger and Chatzisarantis' (2007) and Wilson and Rodgers' (2004) proposals that autonomous motives predict intentions, which, in turn, predict behavior. This suggests that autonomous motives compel people to form intentions to participate in future physical activity, as outlined by the theory of planned behavior. It also corroborates Vallerand's (2007) proposal that the effects of distal, context-level autonomous motives toward physical activity predict motivational constructs (i.e., intentions) at the proximal, situational-level.

While the pattern of prediction and proposed motivational sequence in the trans-contextual model was almost identical to the model reported by Hagger et al. (2003), the proposed mediation of the effect of perceived autonomy support in a PE context on autonomous motivation in a leisure-time context by autonomous motivation in PE was not supported. In fact, while perceived autonomy support directly influenced autonomous motives in both contexts, autonomous motives in PE had no effects on any of the variables in the model. One possible reason for this is that the nature of autonomous motives in PE has no direct bearing on autonomous motives in leisure-time in this sample of adolescents from a Greek high school. To speculate, a possible reason for this may be the general lack of endorsement for autonomous motives in school PE relative to autonomous motives in a leisure time contexts. Therefore, while perceived autonomy support seemed to explain variance in the level of autonomous motivation in both contexts, autonomous motives in PE has no motivational impact outside of that context.

The significant direct effect of perceived autonomy support on leisure-time autonomous motives unmediated by autonomous motives in the PE context may also imply that the link between the two contexts (i.e., PE and leisure) is initiated before the formation of motivation in PE. For example, the satisfaction of the basic psychological needs in an autonomy supportive environment may lead to the formation of autonomous motives for leisure, overriding the PE motives. Or the satisfaction of the needs in PE may influence the needs in leisure activities. So far, there is some preliminary evidence regarding the top-down influence of psychological needs on autonomous forms of motivation in leisure-time contexts (Hagger et al., 2006), but this has not been corroborated in PE or trans-contextual studies.

Conclusion, Recommendations for Future Research and Limitations of the Study

The present study provided evidence to support the premises of Hagger et al.'s (2003) trans-contextual model in a sample of adolescents from a Greek high school. Importantly, the motivational sequence in which perceived autonomy support in PE influences autonomous motives, the decision-making constructs from the theory of planned behavior, and physical activity behavior in a leisure-time context was supported. This suggests that the fostering of autonomous motives in PE may contribute young people's decisions to participate in physical activity outside of school.

The study is not without its limitations, however. It is important to note that some of the constructs such as subjective norms did not exhibit sound psychometric integrity given the low Cronbach alpha levels. While these have been reported in the same measures adopted elsewhere, it does limit the inferences made using this construct in the present study. In addition, the final sample size was relatively small. Given the complex and time-demanding prospective design of the study, such attrition over three waves is not surprising and is a feature of such research. However, we also conducted a check using a bootstrap procedure (Efron, 1982) to support the robustness of the findings presented here. The bootstrap resampling plan generates simulated samples based on the original data and tests the hypothesized model against these additional data sets. Such a plan has been used by previous researchers using complex prospective designs to provide additional support for their findings (Standage et al., 2003). Across many replications (N = 999), the final model from the present study exhibited robust goodness-of-fit statistics with an acceptable range (M CFI = .95; 95% Confidence Interval Lower Bound = .91; 95% Confidence Interval Upper Bound = 1.00). This analysis provides strong evidence that the present model represents a robust description of the data in this sample.

Present results have some important practical implications for PE teachers and those devising school curricula. Such practitioners would do well to adopt the autonomy-supportive behaviors identified by Reeve (2002) to foster autonomy in PE. These include adopting a questioning approach, providing a choice of activity, acknowledging conflict in decision-making, setting personally-relevant goals, and providing a clear rationale for the activity. These results also advance knowledge in the field by replicating the trans-contextual model in a culture that reports much higher collectivist values compared with the culture in which the model was originally developed. It seems the pattern of effects in the model may be universal as indicated by this and other cross-cultural studies using the trans-contextual model (e.g., Hagger & Chatzisarantis, 2007). However, future studies may focus on establishing the generalizability of the model further by testing its hypotheses in more diverse cultures, such as those from Asia, which are typically considered to be predominantly collectivist in their cultural orientation. Additionally, while the present study indicated that autonomy support from one important source, that of PE teachers, seems to affect motivation and physical activity behavior in a leisure time context, the present study did not investigate autonomy support from other salient sources such as parents and peers. Future studies will need to account for these sources of autonomy support to provide a more complete model of motivation. Finally, future studies should examine the role of basic psychological needs in the formation of autonomous motives across different contexts (see Hagger et al., 2006).

Authors' Note

The present study was part of a larger study into the cross-cultural applicability of the trans-contextual model

References

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. (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.

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.

Bagozzi, R. P., Lee, H. M., & Van Loo, M. F. (2001). Decisions to donate bone marrow: The role of attitudes and subjective norms across cultures. Psychology and Health, 16, 29-56.

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.

Biddle, S. J. H., Gorely, T., & Stensel, D. J. (2004). Health-enhancing physical activity and sedentary behaviour in children and adolescents. Journal of Sport Sciences, 22, 679-701.

Brislin, R. W. (1986). The wording and translation of research instruments. In W. J. Lonner & J. W. Berry (Eds.), Field methods in educational research (pp. 137-164). Newbury Park, CA: Sage.

Calfas, C. J., & Taylor, W. C. (1994). Effects of physical activity on psychological variables of children. Pediatric Exercise Science, 6, 406-423.

Centers for Disease Control and Prevention. (2003). Physical activity levels among children aged 9-13 years--United States, 2002. Mortality and Morbidity Weekly, 52, 785-788.

Chatzisarantis, N. L. D., & Biddle, S. J. H. (1998). Functional significance of psychological variables that are included in the theory of planned behaviour: 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., 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.

Chirkov, V. I., & Ryan, R. M. (2001). Parent and teacher autonomy support in Russian and U.S. adolescents: Common effects on well-being and academic motivation. Journal of Cross-Cultural Psychology, 32, 618-635.

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, 14291464.

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.

Deci, E. L., & Ryan, R. M. (2003). The Sport Climate Questionnaire. Retrieved February 20, 2003, from Rochester University, Self-Determination Theory Faculty web site: http://www.psych.rochester.edu/SDT/.

Efron, B. (1982). The jackknife, bootstrap and other resampling plans. Philadelphia: SIAM.

Fox, K. R., & Biddle, S. (1989). The child's perspective in P.E. Part IV: Psychology and professional issues. British Journal of Physical Education, 20, 35-38.

Gilliam, T. B., & MacConnie, S. E. (1985). Coronary heart disease risk in children and their physical activity patterns. In R. A. Boileau (Ed.), Advances in Pediatric Sports Sciences (pp. 171-187). Champaign, II.: Human Kinetics.

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.

Guay, F., Mageau, G. A., & Vallerand, R. J. (2003). On the hierarchical structure of self-determined motivation: A test of top-down, bottom-up, reciprocal, and horizontal effects. Personality and Social Psychology Bulletin, 29, 992-1004.

Hagger, M. S., & Armitage, C. (2004). The influence of perceived loci of control and causality in the theory of planned behavior in a leisure-time exercise context. Journal of Applied Biobehavioral Research, 9, 45-64.

Hagger, M. S., Chatzisarantis, N., & Biddle, S. J. H. (2001). The influence of self-efficacy and past behaviour on the physical activity intentions of young people. Journal of Sports Sciences, 19, 711-725.

Hagger, M. S., Chatzisarantis, N., & 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., & 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. (2007). The trans-contextual model of motivation. In M. S. Hagger & N. L. D. Chatzisarantis (Eds.), Intrinsic Motivation and Self-determination in Exercise and Sport (pp. 53-70). Champaign, Il: Human Kinetics.

Hagger, M. S., & Chatzisarantis, N. L. D. (2008). 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., 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.

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.

Jacobs, D. R. J., Ainsworth, B. E., Hartman, T. J., & Leon, A. S. (1993). A simultaneous evaluation of 10 commonly used physical activity questionnaires. Medicine and Science in Sports and Exercise, 25, 92-98.

Mullah, 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. (2005). A prospective study of participation in optional school physical education based on self-determination theory. Journal of Educational Psychology, 97, 444-453.

Oyserman, D., Coon, H. M., & Kemmelmeier, M. (2002). Rethinking individualism and collectivism: Evaluation of theoretical assumptions and meta-analyses. Psychological Bulletin, 128, 3-72.

Reeve, J. (2002). Self-determination theory applied to educational settings. In E. L. Deci & R. M. Ryan (Eds.), Handbook of self-determination research (pp. 183-203). Rochester, NY: University of Rochester Press.

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.

Sallis, J. F., & Patrick, K. (1994). Physical activity guidelines for adolescents: Consensus statement. Pediatric Exercise Science, 6, 302-314.

Sheeran, P., Norman, P., & Orbell, S. (1999). Evidence that intentions based on attitudes better predict behaviour than intentions based on subjective norms. European Journal of Social Psychology, 29, 403-406.

Sheeran, P., Orbell, S., & Trafimow, D. (1999). Does the temporal stability of behavioral intentions moderate intention-behavior and past behavior-future behavior relations? Personality and Social Psychology Bulletin, 25, 721-730.

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.

Triandis, H. C. (1995). Individualism and collectivism. Boulder, CO: Westview 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, I1: Human Kinetics.

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.

Vassilis Barkoukis

Aristotle University of Thessaloniki, Greece

Martin S. Hagger

University of Nottingham, UK

(1) According to Vallerand (2007) a further construct is located beyond the external extreme on the PLOC continuum, amotivation, which reflects the absence of motivation.

(2) Time constraints meant that only a sub-sample of participants were able to complete the Triandis measure.

Address Correspondence To: Vassilis Barkoukis, Department of Physical Education and Sport Science, Aristotle University of Thessaloniki, 54124 Thessaloniki, Greece, Email: bark@phed.auth.gr.

Table 1. Descriptive Statistics and Intercorrelations Among the
Trans-Contextual Model Components

Variable Mean SD [alpha]

1. Perceived Autonomy Support (PE) 4.91 0.82 .82
2. Relative Autonomy Index (PE) 12.36 2.79 .67 (a)
3. Relative Autonomy Index (LT) 19.60 4.30 .65 (a)
4. Attitudes 5.54 1.39 .92
5. Subjective Norm 4.23 1.48 .48
6. Perceived Behavioral Control 5.10 1.27 .69
7. Intention 4.31 1.61 .73
8. Leisure Time Exercise Behavior 3.33 1.68 --

Variable 1 2 3

1. Perceived Autonomy Support (PE) --
2. Relative Autonomy Index (PE) .24 * --
3. Relative Autonomy Index (LT) .42 ** .24 * --
4. Attitudes .38 ** .17 .60 **
5. Subjective Norm .19 .13 .31 **
6. Perceived Behavioral Control .31 * .11 .56 **
7. Intention .23 ** .02 .45 **
8. Leisure Time Exercise Behavior .23 * -.01 .41 **

Variable 4 5 6 7

1. Perceived Autonomy Support (PE)
2. Relative Autonomy Index (PE)
3. Relative Autonomy Index (LT)
4. Attitudes --
5. Subjective Norm .22 * --
6. Perceived Behavioral Control .55 ** .29 ** --
7. Intention .60 ** .18 .58 ** --
8. Leisure Time Exercise Behavior .41 ** .19 .39 ** .45 **

Note. PE = PE Context; LT = Leisure Time Context (a) Coefficient
represents the median alpha coefficient for the intrinsic motivation,
identified regulation, introjected regulation, and external regulation
constructs that form the relative autonomy index composite

* p < .05. ** p < .01.