Table 2.
Direct, indirect, and total associations.
| Predictor and outcome | Direct β | 95% BC CI | Total indirect β | 95% BC CI | Total β | 95% BC CI |
|---|---|---|---|---|---|---|
| (A) Selected latent-variable direct, total indirect, and total associations estimated in Amos | ||||||
| PAI and IA | 0.172 | [0.112, 0.233] | 0.289 | [0.245, 0.336] | 0.461 | [0.399, 0.517] |
| PAI and SRL | 0.240 | [0.182, 0.298] | −0.382 | [−0.446, −0.317] | −0.142 | [−0.210, −0.072] |
| IO and SRL | −0.169 | [−0.238, −0.096] | −0.399 | [−0.452, −0.349] | −0.568 | [−0.633, −0.504] |
| PC and SRL | 0.508 | [0.446, 0.568] | 0.095 | [0.070, 0.124] | 0.604 | [0.545, 0.659] |
| Specific indirect path | B | BootSE | 95% percentile bootstrap CI | Fully standardized effect |
|---|---|---|---|---|
| (B) Complementary specific indirect associations estimated from composite scores | ||||
| PAI → IO→SRL | −0.082 | 0.017 | [−0.116, −0.050] | −0.080 |
| PAI → PC → SRL | −0.045 | 0.018 | [−0.081, −0.013] | −0.044 |
| PAI → IA → SRL | −0.056 | 0.011 | [−0.079, −0.036] | −0.054 |
| PAI → IO→PC → SRL | −0.105 | 0.012 | [−0.129, −0.083] | −0.102 |
| PAI → IO→IA → SRL | −0.052 | 0.008 | [−0.069, −0.038] | −0.051 |
| PAI → PC → IA → SRL | −0.009 | 0.004 | [−0.017, −0.003] | −0.009 |
| PAI → IO→PC → IA → SRL | −0.021 | 0.004 | [−0.029, −0.015] | −0.021 |
| Total indirect | −0.371 | 0.032 | [−0.436, −0.310] | −0.360 |
| Direct | 0.23 | 0.029 | [0.173, 0.287] | 0.223 |
| Total | −0.141 | 0.035 | [−0.211, −0.071] | −0.137 |
Panel A reports selected standardized latent-variable direct, total indirect, and total associations estimated in Amos, with 95% bias-corrected bootstrap confidence intervals based on 5,000 resamples. Panel B reports ordinary least squares estimates based on composite scale scores. Within Panel B, the seven specific indirect estimates sum, subject to rounding, to the composite-score total indirect association. The latent-variable total indirect associations in Panel A are estimated separately. Bootstrap intervals in Panel B are percentile intervals based on 5,000 samples.