4.2. Assessment of Measurement Model: Reliability and Validity Analysis
In this study, Cronbach's α test and confirmatory factor analysis (CFA) were conducted to verify the reliability and validity of the measurement tools used. Particularly, as the reliability and validity of the measurement tool can vary depending on the subject, this study not only considered the entire group but also divided the group into those with high and low eco-friendly attitudes to verify whether the measured tools for each group were at a satisfactory level. For eco-friendly attitudes, the average value was 5.11, and the group was divided (mean split) based on this average.
Confirmatory factor analysis was conducted to analyze the fit of the measurement model. The results showed that the entire group (χ2 = 198.822[df = 105, p<.01], χ2/df = 1.893, TLI = .927, CFI = .944, RMSEA = .058, SRMR = .062), the group with high eco-friendly attitudes (χ2 = 211.454[df = 105<.01], χ2/df = 2.013, TLI = .897, CFI = .881, RMSEA = .062, SRMR = .069), and the group with low eco-friendly attitudes (χ2 = 231.778[df = 105<.01], χ2/df = 2.207, TLI = .847, CFI = .834, RMSEA = .064, SRMR = .067) all exhibited satisfactory levels.
Upon conducting the reliability analysis, the reliability α values of each sub-domain for the entire group were between .856–.917, for the group with high eco-friendly attitudes they were between .856–.911, and for the group with low eco-friendly attitudes, they were between .877–.914. Since a reliability level of .6 or above is considered satisfactory, it was confirmed that the internal reliability of the items used in this study were all at a trustworthy level.
Table 5.
Reliability Analysis of Measurement Items (CFA).
Table 5.
Reliability Analysis of Measurement Items (CFA).
Variable |
b |
SE |
t |
α |
CR |
AVE |
Total |
EA(H) |
EA(L) |
Total |
EA(H) |
EA(L) |
Total |
EA(H) |
EA(L) |
AUT 1 |
1.000 |
- |
- |
.917 |
.911 |
.924 |
.950 |
.898 |
.778 |
.863 |
.747 |
.540 |
AUT 2 |
.981 |
.031 |
21.06 |
|
|
|
|
|
|
|
|
|
AUT 4 |
.905 |
.033 |
19.55 |
|
|
|
|
|
|
|
|
|
COM 2 |
1.000 |
- |
- |
.909 |
.900 |
.886 |
.875 |
.793 |
.812 |
.701 |
.561 |
.591 |
COM 3 |
.895 |
.066 |
30.01 |
|
|
|
|
|
|
|
|
|
COM 5 |
.756 |
.051 |
17.91 |
|
|
|
|
|
|
|
|
|
REL 1 |
1.000 |
- |
- |
.903 |
.873 |
.910 |
.835 |
.831 |
.807 |
.629 |
.623 |
.584 |
REL 2 |
.815 |
.037 |
18.66 |
|
|
|
|
|
|
|
|
|
REL 3 |
.701 |
.020 |
18.89 |
|
|
|
|
|
|
|
|
|
MI 1 |
1.000 |
- |
- |
.911 |
.856 |
.896 |
.866 |
.845 |
.799 |
.683 |
.647 |
.572 |
MI 2 |
.853 |
.030 |
20.35 |
|
|
|
|
|
|
|
|
|
MI 3 |
.768 |
.067 |
17.53 |
|
|
|
|
|
|
|
|
|
EI 1 |
1.000 |
- |
- |
.856 |
.889 |
.877 |
.752 |
.812 |
.784 |
.504 |
.591 |
.548 |
EI 2 |
.749 |
.056 |
15.64 |
|
|
|
|
|
|
|
|
|
EI 3 |
.657 |
.075 |
16.35 |
|
|
|
|
|
|
|
|
|
The CR value for verifying the convergent validity of the measurement tool was .752–.950 for the entire group, .793–.898 for the group with high eco-friendly attitudes, and .778–.812 for the group with low eco-friendly attitudes. Considering that the commonly accepted figure is .5, it was confirmed that the measurement tool of this study is at an appropriate level [
58]. The AVE value, which was confirmed to verify the discriminant validity of the measurement tool, was .504–.863 for the entire group, .561–747 for the group with high eco-friendly attitudes, and .540–591 for the group with low eco-friendly attitudes. Considering that the commonly accepted figure is between .5–.95, it was confirmed that the measurement tool of this study will secure discriminant validity [
59].
4.3. The Role of Motivation in Mediating the Relationship between Self-Determination Factors and Sustained Participation Intention: Structural Equation Analysis (Study 1)
This study conducted a structural equation modeling analysis to examine the relationship between the degree of self-determination factors experienced by plogging participants, their intrinsic and extrinsic motivation, and their intention to continue participating. Based on the hypotheses of research question 1 (Hypotheses 1-1 to 3-3), we constructed and analyzed the research model. First, the fit of the research model was analyzed and the results showed that X2(df)= 479.08(219), TLI=.915, CFI=.903, and RMSEA= .056. This indicates that the structural equation model is at an acceptable level for interpretation [
60].
Looking at the path process of the structural model presented in [
Figure 1] in detail, the basic psychological needs factors of autonomy (b=.422, p<.001), competence (b=.214, p<.05), and relatedness (b=.681, p<.001), proposed in the Self-Determination Theory, were found to have a positive (+) impact on intrinsic motivation. This suggests that the stronger the experience of autonomy and competence in past plogging participation and the better the relatedness with the people who participated together, the stronger the intrinsic motivation for plogging participation. Among these, the experience of relatedness had the greatest impact on intrinsic motivation, and the experience of competence was analyzed as the lowest. Conversely, extrinsic motivation showed a somewhat different trend, where the intensity of the experience of autonomy (b=.155, p>.05) and competence (b=.021, p>.05) did not have a significant impact, and the experience of relatedness with the people who participated together (b=.245, p<.01) had a significantly positive (+) impact. This suggests that the relationship between the formation of plogging exercise participation motivation, according to the type and degree of basic human needs satisfaction proposed in the Self-Determination Theory, is mainly related to intrinsic motivation, and for extrinsic motivation, only relatedness has an impact.
In contrast, when analyzing the direct effect of the basic psychological needs satisfaction factors proposed in the Self-Determination Theory on the relationship with the intention to continue plogging participation, the experience of autonomy (b=.004, p>.05) and competence (b=.098, p>.05) did not have a significant impact, whereas the experience of relatedness with the people who participated together (b=.466, p<.001) had a direct positive (+) effect. In addition, the intrinsic motivation (b=.832, p<.001) and extrinsic motivation (b=.089, p<.05) set as mediators both had a statistically significant positive (+) impact on the intention to continue participating, with intrinsic motivation having the greater impact.
In addition, an effect decomposition analysis was conducted to analyze the effect of intrinsic and extrinsic motivation mediating the relationship between the basic psychological needs satisfaction experience through plogging activities and the intention to continue plogging participation. The analysis results are presented in [
Table 6].
First, in the case of autonomy, the direct effect on the intention to continue plogging participation and extrinsic motivation was not significant, and only a direct effect on intrinsic motivation was observed. Furthermore, since intrinsic motivation had a direct impact on the intention to continue participating, the mediation model of intrinsic and extrinsic motivation mediating the impact of autonomy on the intention to continue participating did not hold because the direct effect of autonomy on the intention to continue participating was not significant (b=.004, p>.05). Similarly, in the case of competence, the direct effect on the intention to continue participating was not significant, so the mediation model did not hold (b=.098, p>.05). In contrast, in the case of relatedness, the direct effect on the intention to continue participating was significant (b=.466, p<.001), and the effects of relatedness on both intrinsic motivation (b=.681, p<.001) and extrinsic motivation (b=.242, p<.05) were significant. Since both intrinsic motivation (b=.832, p<.001) and extrinsic motivation (b=.089, p<.05) had significant effects on the intention to continue participating, the mediation model held. Specifically, the indirect effect of intrinsic motivation mediating the relationship between relatedness and the intention to continue participating was .566 (B.C: .178~.397, p<.001), and the indirect effect of extrinsic motivation was .021 (B.C: .012~.201, p<.001). The indirect effect of intrinsic motivation was greater than the direct effect of relatedness on the intention to continue participating, and the indirect effect of extrinsic motivation was smaller. The sum of the indirect effects of intrinsic and extrinsic motivation (.587) was greater than the direct effect of relatedness on the intention to continue participating (.466), indicating that intrinsic and extrinsic motivation fully mediate the relationship between relatedness and the intention to continue participating. However, the fact that the indirect effect of extrinsic motivation (.021) was significantly smaller than that of intrinsic motivation (.566) confirms the important role of intrinsic motivation.
4.4. Verification of Group Differences According to Eco-friendly Attitudes: Multi-group Structural Equation Analysis (Study 2)
This study performed a multi-group structural equation analysis to verify the differences in motivation, which mediates the relationship between the degree of perceived basic psychological needs satisfaction and the intention to continue participation through plogging participation analyzed earlier, according to eco-friendly attitudes. The average value of eco-friendly attitudes was 5.11, and the group with a lower average was named the low eco-friendly attitude group (146 people), and the group with a higher average was named the high eco-friendly attitude group (142 people).
To compare the path coefficients between the two groups, we analyzed the similarity between the two models by comparing the form similarity and structure similarity. The results showed that the unconstrained model and the form similarity model exhibited a good fit with χ2= 524.375(df=187), CFI= .915, and RMSEA= .051. In the measurement similarity model, when the path between the latent variable and the measurement variable was constrained to be the same, χ2=589.758(df=200), CFI= .913, and RMSEA= .048, the model and data showed a relatively good fit and there was no statistically significant difference from the form similarity model. Through this, it was confirmed that the observed variables measuring each construct were perceived equally between groups. In addition, in the structural similarity model, when the variance and covariance of the latent variables were constrained, it was analyzed as χ2=605.217(df=218), CFI= .910, and RMSEA= .044, and the model and data showed a relatively good fit. Furthermore, there was a statistically significant difference in the χ2 difference comparison between the form similarity and measurement similarity models. Based on these results, it was confirmed that the two groups, divided according to eco-friendly attitudes, have an effect as control variables, and a path coefficient comparison analysis was performed for the high and low groups of eco-friendly attitudes. The analysis results are shown in [
Table 7].
First, in the case of the influence of autonomy on intrinsic motivation, both the low eco-friendly attitude group (b=.402, p<.001) and the high group (b=.500, p<.001) had a significant positive influence, and no difference was found in the two path coefficients (∆χ2= .94, p>.05). In the case of the influence of autonomy on extrinsic motivation, while the low eco-friendly attitude group had a significant influence (b=.243, p<.05), the high eco-friendly attitude group did not have a statistically significant influence (b=.113, p>.05). The analysis of the difference in path coefficients was statistically significant (∆χ2= 3.09, p<.01). In the case of the influence of autonomy on the intention to continue participation, neither the high eco-friendly attitude group (b=.073, p>.05) nor the low group (b=.085, p>.05) had a significant influence, and the difference in the two path coefficients was also not significant (∆χ2= .55, p>.05). In the case of the influence of competence on intrinsic motivation, while the low eco-friendly attitude group (b=.194, p<.01) was significant, the high eco-friendly attitude group (b=.054, p>.05) was not significant, and the difference in the two path coefficients was also significant (∆χ2= .4.01). In the case of the influence of competence on extrinsic motivation, neither the low eco-friendly attitude group (b=.001, p>.05) nor the high group (b=.056, p>.05) had a significant influence, and no difference was found between the two path coefficients (∆χ2= .080, p>.05). In the case of relatedness, both the low eco-friendly attitude group (b=.586, P<.001) and the high eco-friendly group (b=.633, P<.001) had a significant influence on intrinsic motivation. Both the low eco-friendly attitude group (b=.563, P<.001) and the high eco-friendly attitude group (b=.412, P<.001) had a significant influence on the intention to continue participation. However, only the high eco-friendly attitude group (b=.328, P<.05) had a significant influence on extrinsic motivation. The path coefficient difference verification results showed no significant difference between the two groups in the influence on intrinsic motivation (∆χ2= .98, p>.05) and the influence on the intention to continue participation (∆χ2= .43, p>.05), but a significant difference was confirmed between the two groups in the case of extrinsic motivation (∆χ2= 2.79, p<.01).
To understand the different patterns of mediating effects that can occur depending on eco-friendly attitudes, we compared the mediating effect analysis for each model of the low and high eco-friendly attitude groups. First, in the case of the low eco-friendly attitude group, the direct effect of autonomy on the intention to continue participation was not significant (b=.55, p>.05), so the mediating effects of intrinsic motivation and extrinsic motivation mediating autonomy and the intention to continue participation were not established. The same was true for competence, where the direct effect on the intention to continue participation (b=.027, p>.05) was not significant, so the mediating effects of intrinsic motivation and extrinsic motivation were not established. It was confirmed that there were significant correlations between relatedness and intrinsic motivation (b=.633, p<.001), relatedness and extrinsic motivation (b=.172, p<.05), intrinsic motivation and the intention to continue participation (b=.447, p<.001), and extrinsic motivation and the intention to continue participation (b=.104, p<.05). Also, the relationship between relatedness and the intention to continue participation (b=.412, p<.001) was significant. Through this, it was confirmed that the model in which intrinsic and extrinsic motivation mediate relatedness and the intention to continue participation was established, and the significance of the mediating effect was also confirmed through the Sobel test (B.C:. 012~.131, p<.001). Specifically, it was confirmed that only intrinsic motivation partially mediates the relationship between relatedness and the intention to continue participation, considering that the influence of relatedness on extrinsic motivation was not significant and that the direct effect (b=.563) of relatedness on the intention to continue participation was greater than the indirect effect (b=.244) that influences through intrinsic motivation.
In both the high and low eco-friendly attitude groups, the mediating effects of intrinsic and extrinsic motivation were not established because there was no significant correlation between the intention to continue participation and autonomy and competence. However, a somewhat different pattern of mediating effects was found in relatedness. It was confirmed that there were significant correlations between relatedness and intrinsic motivation (b=.633, p<.001), and between intrinsic motivation and the intention to continue participation (b=.447, p<.001). In addition, the relationship between relatedness and the intention to continue participation (b=.412, p<.001) was significant. However, the relationship between extrinsic motivation and the intention to continue participation (b=.028, p>.05) was not significant. Through this, it was confirmed that the model in which intrinsic motivation mediates relatedness and the intention to continue participation was established, and the significance of the mediating effect was also confirmed through the Sobel test (B.C:. 049~.205, p<.001). Specifically, it was confirmed that only intrinsic motivation partially mediates the relationship between relatedness and intention.