Opponent-sponsored polls understate the candidate by ~2 pp

🟡 β_opp = −1.93 (SE 0.89, p = 0.030) in Spec 2 of AN-001. The symmetry test gives β_self − β_opp ≈ +9.7 pp, supporting the sender-specific reading: bias operates on the sponsor's own candidate rather than as a generic pollster house effect.

Single-source 🟡 because no independent design separately validates the negative direction.

Sources.

Own analysis: build/table/regressions.csv (AN-001 Spec 2 — opponent_sponsored row); AN-001
Cross-refs: H2 opponent-sponsored-mirror; sender-specific interpretation

Cited analyses

AN-001 green causal

Within-candidate FE on 568 self-sponsored candidate-poll rows gives β = +7.75 pp (p<0.001) — large, robust to spec, robust to renormalization choice. The opponent-sponsored coefficient is -1.93 (p=0.030), so the bias is sender-specific not a generic house effect.