Skip to main content
. 2026 Oct 7;17:1937810. doi: 10.3389/fpsyg.2026.1937810

Table 4.

Specific indirect associations in the revised model.

Specified product Estimate Bootstrap SD p 95% percentile CI
EEX → EFD → PC 0.0048 0.0073 0.509 [−0.0089, 0.0201]
RAC → EEX → EFD 0.0018 0.0033 0.587 [−0.0038, 0.0097]
RAC → EFD → PC 0.0309 0.0098 0.002 [0.0137, 0.0517]
RAC → EEX → PC 0.0166 0.0102 0.104 [−0.0035, 0.0371]
TAC → EEX → EFD 0.0019 0.0033 0.561 [−0.0042, 0.0094]
TAC → EFD → PC 0.0455 0.0115 <0.001 [0.0252, 0.0696]
TAC → EEX → PC 0.0176 0.0100 0.078 [−0.0005, 0.0390]
TRC → EEX → EFD 0.0106 0.0157 0.500 [−0.0196, 0.0419]
TRC → EFD → PC 0.0386 0.0121 0.001 [0.0181, 0.0648]
TRC → EEX → PC 0.0965 0.0175 <0.001 [0.0647, 0.1329]
TRC → EEX → EFD → PC 0.0020 0.0031 0.515 [−0.0037, 0.0085]
RAC → EEX → EFD → PC 0.0003 0.0006 0.595 [−0.0007, 0.0019]
TAC → EEX → EFD → PC 0.0004 0.0006 0.575 [−0.0008, 0.0019]

All 13 coefficient products implied by the 12 specified paths are reported, including nine dimension-to-PC products; four intervals exclude zero. Inference uses the prespecified percentile intervals. For TAC–EEX–PC, p = 0.078 and the unrounded lower endpoint is −0.00047095; the exported bias-corrected interval is not used for inference. All three dimension–EEX–EFD–PC intervals include zero. These estimates describe associations within one survey wave, leaving temporal ordering unresolved.