United States Postal Service
vendor · #10 of 310 in Pennsylvania · U.S. Senate · 2024all Schedule E rows for this payee on fec.gov ↗verified by hand · rule case/punctuation/suffix normalisation + token-set ≥0.92 (numeric tokens must match); aliases file — hand alias row folds 3 filed payee string(s) into one vendor — checked by devin-block2-vendors on Sep 5, 2026fec.gov ↗
By medium
Medium is classified from the purpose the spender filed; the FEC does not record which buy placed which ad.
inferred · rule Medium is classified from the purpose string the spender filed on Schedule E, by an ordered keyword table (first match wins, keywords matched at word starts): production: produc, creative, design, videograph, photograph, filming; digital: digital, online, internet, web, social, streaming, email, youtube, facebook, google, meta, gif; radio: radio; tv: tv, television, broadcast, cable, media buy, media placement; mail: mail, postage, postcard, print; phones: phone, text, sms, robocall, call, dial, teleconferenc; consulting: poll, research, consult, strateg, list, targeting, voter file; field: canvass, door, field, gotv; other: no match. The FEC does not record which buy placed which ad.Schedule E fields ↗
Who paid this vendor
| Spender | Paid | Buys |
|---|---|---|
| AMERICANS FOR PROSPERITY ACTION, INC. (AFP ACTION) DBA CVA ACTION AND DBA LIBRE ACTION | $7,191,078 | 35 |
| ACTION TOGETHER NEPA INC. | $1,916 | 1 |
| PENNSYLVANIA PRO-LIFE FEDERATION PAC | $1,280 | 4 |
Candidates named on the buys
Targeting, not money: the spender declares support or opposition on each Schedule E row; the candidate receives nothing.
Ads linked to this vendor
None. An ad is linked only when a person verified it from a source naming both, or when this was the only digital vendor its sponsor paid in the week before and while it ran — the FEC does not record which buy placed which ad, so this is not evidence that the vendor made no ads.
Every buy
Every Schedule E independent expenditure names a payee. Payee strings are normalised (case, punctuation, legal suffixes, token order), near-duplicates are folded with a difflib ratio ≥ 0.92 when their numeric tokens match, and hand alias rows from data/hand/<race>/vendor_aliases.json override the rule. Medium is classified from the purpose the spender filed (ordered keyword table in campaign_commons/vendors.py). The FEC does not record which buy placed which ad: vendor ↔ ad links are date-window adjacency (or single-digital-vendor inference) from campaign_commons.ads_enrich, each with its basis, and are only `verified` when a person found a source naming both. Totals are the filed amounts, grouped, never re-estimated.