Own-CPF sponsorship route gives β ≈ +19 vs +6-9 for committee/party

🟡 When the candidate pays with their own CPF (n_self = 18), β = +19.12 (p = 0.006) — more than twice the +6-9 of the committee/party CNPJ routes. Smallest cell but the sharpest single-route signal.

Two readings:

Strategic-incentives: direct personal stake → more pressure on the pollster.
Selection: CPF-payers are systematically different (smaller budget, possibly without a formal committee, possibly first-time candidates).

Not distinguishable on n = 18; both readings consistent with the direction.

Sources.

Own analysis: build/table/heterogeneity.csv (route_split rows); AN-006
Cross-refs: H9 funding-source-gradient

Cited analyses

AN-006 yellow causal

When the candidate pays with their own CPF (n_self=18), β = +19.12 pp (p=0.006) — roughly 2× the committee/party routes. Smallest cell but the sharpest signal; suggests strategic incentives concentrate where the sponsor has most personal stake.