3. Results
A comparison of the study measures in boys and girls is presented in
Table 1. As expected, LTEQ score, SRHI, PPF and SE mean scores as well as motivation to exercise (RAI from the BREQ-2) were higher in boys than girls. No significant differences were found when contrasting amotivation and introjected regulation with the PLOC-R. In addition, external regulation from the PLOC-R was higher in girls, while the identified and intrinsic types as well as the RAI from the PLOC-R were higher in boys. All differences demonstrated small to medium effect sizes.
Furthermore, the correlations between the PLOC-R subscales, LCQ and the RAI from the BREQ-2 are presented in
Table 2. The score of the PLOC-R amotivation subscale correlated positively with external and introjected regulation, and negatively with identified and intrinsic regulation. Positive correlations were observed between the PLOC-R subscales representing greater autonomous motivation for PE, with the strongest magnitude between the identified and intrinsic regulation of 0.75 (
p < 0.001). Similar trends in the associations were found between the PLOC-R subscales, the LCQ, and the RAI from the BREQ-2: the perceived learning climate and autonomous motivation to exercise correlated positively with more autonomous PLOC-R subscales and negatively with controlled motivation. Introjected regulation from the PLOC-R had a weak positive correlation with the perceived learning climate and negative one with the motivation to exercise.
Table 3 shows the bivariate correlations between the PLOC-R subscales, LCQ and PA, SRHI for PA, frequency of skipping PE classes, PPF and SE. The identified and intrinsic regulation subscale scores from the PLOC-R and the LCQ score exhibited weak to medium positive associations with PA, PA habits, perceived physical fitness and self-esteem. The frequency of skipping PE classes had weak to medium negative relationships with the learning climate and greater autonomous motivation for PE classes subscales. The amotivation and external motivation subscales’ scores from the PLOC-R also demonstrated weak to medium negative relationships with PPF and SE. On the contrary, positive correlations between the greater controlled motivation for PE subscales from the PLOC-R and the frequency of skipping PE classes were found.
Finally, we created a path model based on the SDT. We tested the associations between perceived learning climate, PA habits and frequency of skipping PE classes mediated by motivation for PE. The final path model with the standardized regression weights is presented in
Figure 2. It was revealed that perceived teacher autonomy support during PE classes had direct positive effects on PA habits (estimate = 0.15; 95% CI = 0.08–0.22;
p < 0.001) and motivation for PE (estimate = 0.40; 96% CI = 0.39–0.45;
p < 0.001), a negative one on the frequency of skipping PE classes (estimate = −0.14; 95% CI = −0.20–(−0.07);
p = 0.001). Furthermore, there were direct effects from PE motivation to PA habits (estimate = 0.11; 95% CI = 0.04–0.20,
p = 0.026) and the frequency of skipping PE classes (estimate = −0.24; 95% CI = −0.30–(−0.18);
p < 0.001). The model demonstrated good fit indices (CFI = 0.998; RMSEA = 0.023 (90% CI = 0.00–0.11); SRMR = 0.012). In addition, configural invariance of the final model across gender groups was tested. Results showed that the model fitted the data across both boys and girls, χ
2 = 7.526,
p = 0.376; df = 7; CFI = 0.998; SRMR = 0.030; RMSEA = 0.015 (90% CI = 0.000, 0.068), and the chi-square test difference between unconstrained and constrained models was not significant (
p = 0.21).
4. Discussion
In this study, we investigated the associations between teacher autonomy support, self-determined motivation for PE, the strength of PA habits and non-participation in PE classes in a sample of adolescents. Based on the SDT, we tested a hypothetical model and expected that teacher autonomy support would be directly and positively associated with the strength of PA habits and negatively with non-participation in PE classes. Furthermore, we hypothesized that greater autonomous motivation for PE would be a mediator between teacher autonomy support and PA habits and non-participation in PE. The results confirmed our assumptions and are in line with the main assumptions of SDT [
11]. The findings of the present study extend the literature on this topic and provide important new data suggesting that enhanced autonomous motivation in PE context is associated with increased stable and persistent PA habits outside the school. Previous works that applied SDT to PE confirmed that social context and autonomous motivation are associated with positive PE outcomes outside of school. Specifically, autonomous motivation for PE is associated with a higher intention to be physically active and actual reported and objectively measured PA during recess and outside of school [
15,
18,
19,
22,
29,
30,
73,
74,
75]. This research adds important knowledge and goes in line with the knowledge that higher internal rewards (or intrinsic motivation) are associated with the greater habitual behaviour such as PA [
44,
45]. This is also consistent with previous findings that more autonomous motivation is associated with higher PA habits in adults [
46,
47]. However, this analysis reports these associations in adolescents and the PE context.
PE teachers greatly impact classroom experiences by fulfilling students’ need for autonomy [
8]. Previous studies demonstrated that teacher autonomy support is associated with students’ higher autonomous motivation for PA in leisure time outside of school [
21,
76] and the perceived controlling behaviour of PE teachers was associated with lower levels of intention to be physically active and lower reported PA outside the school [
20,
31]. The present study adds new knowledge that perceived autonomy support in social context (teacher autonomy support for PE) is directly associated with behaviour in different context such as habitual PA behaviour outside the school (PA habits). However, the present study is cross-sectional, and the directions of associations are bidirectional, thus we were not able to conclude whether teacher autonomy support and greater autonomous motivation is the cause of higher PA habits and attendance in PE classes. It might be that students with more athletic identities or those already participating in formal or non-formal sports find more enjoyment and fun in PE and perceive that they receive more support from their PE teachers. In contrast, it might be that students who are more sedentary and less gifted in sport feel more controlled by their teachers and find less pleasure, enjoyment and value in PE, and therefore, they do not develop PA habits and report higher levels of non-participation in PE. The nature of this study does not allow us to answer these questions and future ones in other designs should be implemented to further test these assumptions.
The findings of this study are consistent with previous analyses reporting that self-determined motivation for PE is associated with higher engagement in PE and vice versa [
54]. This investigation provides a new understanding that lower teacher autonomy support is directly associated with non-participation in PE classes and motivation for PE mediates this association. Thus, it is important to promote teacher autonomy support and self-determined motivation for PE when trying to achieve increased class attendance in PE.
Additionally, in the present study, we hypothesized that autonomy support and self-determined motivation for PE would be associated with greater perceived physical fitness and self-esteem in adolescents. The results of the present study partially supported the hypothesis. Teacher autonomy support and self-determined motivation was associated with higher self-esteem and this finding and go in line with findings of the previous studies [
25]. However, perceived physical fitness was not related with teacher autonomy support, but it was associated with more autonomous forms of motivation for PE. This finding overlaps results of previous study in children that showed perceived physical fitness levels were associated with the greater autonomous motivation [
77].
The gender analysis showed that boys’ PA, perceived physical fitness, PA habits and self-esteem were higher than girls’. Previous studies reported the same gender tendencies [
8,
66,
78,
79]. Boys reported significantly lower non-participation in PE compared to girls, and internal motivation for PA and PE classes was higher than girls. However, we observed no significant gender differences in perceived teacher autonomy support in this sample. This contradicts findings from a prior analysis that reported higher perceived teacher autonomy support for adolescent boys compared with girls [
80]. Overall, the results of this investigation suggest that significant differences in PA and internal motivation exist between the genders and boys are more involved and motivated compared to adolescent girls. The reasons why adolescent girls are less motivated for PE and PA might be associated with multiple factors. However, one of the strongest is body image concerns [
81]. Body image concerns, body dissatisfaction, low self-esteem are related to higher external exercise motivation [
82] and this might be one of the reasons why girls express less autonomous PE and PA motivation. Thus, it is important to pay special attention to adolescent girls and implement specific girls-oriented techniques aiming to foster their enjoyment and inherent interest, as well as positive and functional body image and body satisfaction, with the goal of increasing their internal motivation for PE and strengthening their PA habits [
83,
84].
One of the goals of this investigation was the translation and testing of the main psychometric properties of Lithuanian versions of the LCQ [
57] and PLOC-R [
58]. The exploratory factor structure of the LCQ confirmed the one-factor structure and CFA demonstrated satisfactory one-factor fit indices. Furthermore, based on SDT, all the associations between LCQ and other study variables followed the expected directions, confirming the concurrent, discriminant and nomological validity of the instrument. Finally, the internal consistency of the questionnaire was high (Cronbach’s α 0.96). Thus, we can conclude that the Lithuanian version of LCQ is valid and suitable for future studies in adolescent samples. Previous works confirmed the acceptable psychometric properties of the short version of the LCQ [
85]. However, to the best of our knowledge, no in-depth psychometric evidence for the full version of the LCQ is available for testing the LCQ after it has been translated to other languages, meaning that comparison is limited.
However, in this sample the original factor structure of the Lithuanian PLOC-R version was not confirmed. EFA revealed the three-factor solution instead of the five-factor version, as in the original structure of the PLOC-R. Identified and intrinsic regulations formed the first factor, items concerning amotivation and external regulation subscales shaped the second factor and introjected regulation formed a separate third factor. The PLOC-R was previously validated in the German and French languages and the original structure of the questionnaire was replicated [
86]. Nevertheless, in this analysis, the results of the confirmatory factor structure showed an acceptable three-factor structure. Finally, the full gender invariance of the three-factor structure was not confirmed, so the use of the Lithuanian version of PLOC-R comparing genders is limited. The analysis of the internal consistency of the subscales of the PLOC-R revealed that the external regulation subscale has a Cronbach’s α (0.63), which is too low, and that the Cronbach’s α of introjected regulation meets the minimum requirements for the confirmation of internal consistency (0.74). Therefore, in this study, we only used the RAI index from the PLOC-R, but avoided deeper analyses using subscales of the PLOC-R in the structural mediation model. In this analysis, we were not able to test the effect of separate motivational regulations on the outcomes. However, the correlation analysis of the RAI index from the PLOC-R revealed that associations between the PLOC-R and other variables follow the theory-driven directions. Thus, this might be considered a limitation of this study and future ones using a larger adolescent sample should further test the Lithuanian version of the PLOC-R.
Discussing other limitations of the present study, it is important to address the cross-sectional nature of the research that was discussed previously. Further, we did not assess the basic psychological needs satisfaction of students, so we were not able to measure the full motivational sequence (social context, need satisfaction, motivation and outcomes) proposed by SDT when testing our assumptions [
8]. Future works could address this issue. Finally, the generalization of the results of this investigation should be limited, since the sample in this analysis was from Eastern Europe. Future studies should test our findings in other samples.
Finally, based on SDT, future studies might benefit from testing the impacts of basic psychological needs supported PE environments, intrapersonal satisfaction of basic psychological needs and self-determined motivation in PE on PA habit strength and PA habit development in adolescents of various ages. The role of motivation, internal and external PA goals and positive internal reinforcement for the development of PA habits is still not fully understood [
44,
45]. The results of the present study support the transferability of motivation between PE and other contexts in PA confirming trans-contextual modality of motivation. However, seeking to deeper understand mechanisms of habit formation considering TCM and HMIEM theories might be also beneficial for PE practice and science. The roles of situational, contextual and global factors and these types of motivation on outcomes such as the development of PA habits also need exploration. Future studies of other than cross-sectional designs are necessary to explore these questions.
The present study has important practical implications. The results of this investigation can inform PE teachers’ practice by showing that supporting students’ autonomy and self-determined motivation in PE can facilitate increased participation in PE classes and the formation of students’ PA habits. Further, the results of the present cross-sectional study suggest that adolescents with low PA habits, especially girls, might benefit from increased teacher support for autonomy in PE and increased intrapersonal PE and PA motivation. The outcomes of this examination might also contribute towards PA promotion interventions for adolescents. Providing more intensive autonomy support for adolescents with low PA habits might be an effective strategy in intervention programs aiming to promote healthy lifestyle and long-lasting PA in adolescents.