PHILIA-EcoSensory Swarm Phase SC-9a Pre-registration v0.5: Synthetic Sub-Family Decomposition and MSE×Sub-Class Interaction Test — AR vs Chaos AC1 Dose-Response and OLS Interaction, v32 Apparatus, Independent Seeds 72–76

This document constitutes the pre-registration protocol for PHILIA Phase SC-9a, the confirmatory analysis stage of the SC-9a pipeline. SC-9a applies the locked v32 self-raw × 3-channel normalization apparatus to the 182-substrate integrated registry (SHA256: 3948b55114a6bad44296fa78cb6d53f8fc89d553137d3e6c0d683023344f7c9a) using five independent seeds 72–76, totaling 910 runs (182 substrates × 5 seeds × T=80, 000 steps). Two confirmatory hypotheses are pre-registered under Bonferroni correction (αfamily=0. 05, k=2, αₚerₜest=0. 025). H-SC9a-Primary is a device-integrity replication control: AR sub-family (n=19, AC1<0. 95), Spearman ρ (θᵣank, ρₐt_θ) ≤ −0. 90, exact permutation p<0. 025, 6 θ-levels 0. 10, 0. 30, 0. 50, 0. 70, 0. 80, 0. 95. Type I error P (ρ≤−0. 90|H₀, n=6) = 0. 0083 ✅. No escape clause. H-SC9a-Secondary tests whether the MSE→z slope differs between AR and Chaos Synthetic sub-families via OLS interaction model (z ~ MSE + AC1 + C (subclass) + MSE×C (subclass), AR+Chaos N=40). Verdict: interaction p<0. 025 (single criterion; partial η² reported as effect size standard — conjoint redundancy disclosed). Pre-declared power: 0. 42 (analytical) / 0. 32 (Monte Carlo) at η²=0. 10 — SC-8 L-8 continuity. Pre-registration lineage: SC-8 Results Report (10. 5281/zenodo. 20621215) → SC-9a Pre-registration (this document). SC-9b (Financial ARFIMA + LIGO H1/L1) maintained as separate track. Gate-0 v10 Cycle 9: UNCONDITIONAL ACCEPT (크리스 / Claude Opus / Anthropic). Power analysis SHA256: 35002523431384739d135f60b666364485c450c0325a5814cbd53f9d6a588a39 Trinity AI Research Team | Konyang University | 2026. 06. 11Measurement is life. Description, not proof. Pre-registration protects us. — PHILIA OS | 0∞1∞0. 5∞