Results: Structural Model
To test the proposed framework, a structural equation model was specified with three latent constructs: coping with stress, self-regulation and subjective well-being. Each construct was represented by several observed subscales, and two background variables (age and number of siblings) were entered as exogenous predictors. This specification made it possible to estimate direct and indirect paths simultaneously and, in a second step, to examine whether the structural paths differ between groups.
Parameters were estimated by maximum likelihood (NLMINB optimizer) on 318 cases with 65 free parameters; the solution converged normally after 121 iterations.
Global fit was mixed. The chi-square test was significant (χ² = 642, df = 144, p < .001), and the RMSEA of 0.104 (90% CI 0.096 to 0.112) lies above the conventional 0.08 cut-off. Incremental indices were likewise below the recommended 0.90 (CFI = 0.812, TLI = 0.776), whereas the SRMR of 0.074 indicates an acceptable fit. Taken together, the indices suggest that the model captures the main structure of the data but would benefit from respecification.
Turning to the structural paths, age was negatively related to coping (Estimate = -0.12, β = -0.31, p < .001), while the number of siblings had no effect (p = .668). Self-regulation emerged as the strongest predictor of coping (Estimate = 0.91, β = 0.94, p < .001). Age was positively related to self-regulation (Estimate = 0.31, β = 0.71, p < .001), and siblings again showed no effect (p = .597). None of the predictors had a significant direct effect on well-being (age p = .230, siblings p = .739, coping p = .359). The pattern therefore places self-regulation at the centre of the coping process, with age shaping both constructs.
Table 4. Structural parameter estimates
| Dependent | Predictor | Estimate | SE | β | 95% CI (β) | z | p | |
|---|---|---|---|---|---|---|---|---|
| Lower | Upper | |||||||
| Coping | Age | -0.118 | 0.027 | -0.310 | -0.449 | -0.171 | -4.370 | < .001 |
| Coping | Siblings | 0.006 | 0.014 | 0.020 | -0.071 | 0.111 | 0.429 | .668 |
| Coping | Self-regulation | 0.912 | 0.088 | 0.940 | 0.762 | 1.118 | 10.364 | < .001 |
| Coping | Well-being | 0.054 | 0.390 | 0.006 | -0.079 | 0.091 | 0.138 | .890 |
| Self-regulation | Age | 0.312 | 0.024 | 0.710 | 0.603 | 0.817 | 13.000 | < .001 |
| Self-regulation | Siblings | -0.009 | 0.017 | -0.020 | -0.094 | 0.054 | -0.529 | .597 |
| Well-being | Age | -0.006 | 0.005 | -0.110 | -0.290 | 0.070 | -1.200 | .230 |
| Well-being | Siblings | -0.001 | 0.003 | -0.030 | -0.206 | 0.146 | -0.333 | .739 |
| Well-being | Coping | 0.011 | 0.012 | 0.090 | -0.102 | 0.282 | 0.917 | .359 |
A multi-group model then compared boys and girls using constraint score tests. The omnibus test was significant (χ² = 47.36, df = 23, p = .002), showing that at least some parameters are not equal across groups, mainly those linking self-regulation and coping.
The clearest difference concerned the path from self-regulation to coping (χ² = 6.41, p = .011), suggesting that the extent to which self-regulation translates into effective coping is not the same for boys and girls.