supported 22 · mismatch 0 · not-in-source 19 · wrong-source 0. Report-only; no pages edited.
Three structural patterns account for all 19 P2 findings:
- EFSA 2020 misattribution (6 rows, almond): Flyvholm 1984 values appear in a sentence whose trailing citation is EFSA 2020 (which begins the next sentence about a different claim). The audit tool captures EFSA 2020 as nearest citation; EFSA 2020 source page contains no per-food Ni means.
- FDA TDS source page body gap (9 rows, apple-juice + apple): The
fda2022-tds-elements-fy2018-fy2020source page documents the dataset and references the underlying CSVs but does not enumerate per-food values in its body. All nine values exist indata/evidence/fda_tds_fy2018_2020_summary_by_food_analyte.csv; needs trivial CSV-side verification. - Loose method citation (3 rows, ACV) and one out-of-range approximation (1 row, apple-juice): Ndungu 2004 is cited for detection floors it doesn’t state; Benoy 1971 is cited for a 200 mg/kg threshold that doesn’t appear in its source page.
P2 findings — not-in-source
| Page | Section | Claim | Cited source | Verdict | Note |
|---|---|---|---|---|---|
| almond | Why this commodity accumulates heavy metals | 1.3 µg/g | efsa-nickel-contam-2020 | not-in-source | Value is from Flyvholm 1984 (preceding citation in same sentence). EFSA 2020 citation starts the next sentence (“EFSA’s 2020 Nickel opinion names nuts collectively…”). EFSA 2020 source page contains no almond Ni mean. |
| almond | Why this commodity accumulates heavy metals | 300 ppb | efsa-nickel-contam-2020 | not-in-source | Fragment of “1,300 ppb” (= 1.3 µg/g); same attribution issue as row above. |
| almond | Nickel in almond | 1.8 µg/g | efsa-nickel-contam-2020 | not-in-source | Hazelnut mean Ni from Flyvholm 1984 (Key numbers: “Hazelnuts | 0.66–3.3 | 12 | 1.8”). Sentence containing these values ends with EFSA 2020 citation that governs a different following claim; no citation in EFSA 2020 source page for 1.8 µg/g. |
| almond | Nickel in almond | 3.6 µg/g | efsa-nickel-contam-2020 | not-in-source | Walnut mean Ni from Flyvholm 1984 (Key numbers: “Walnuts | single | 1 | 3.6”). Same attribution issue. |
| almond | Nickel in almond | 2.8 µg/g | efsa-nickel-contam-2020 | not-in-source | Peanut mean Ni from Flyvholm 1984 (Key numbers: “Peanuts | 1.6–4.9 | 2 | 2.8”). Same attribution issue. |
| almond | Nickel in almond | 0.8 µg/g | efsa-nickel-contam-2020 | not-in-source | Pistachio mean Ni from Flyvholm 1984 (Key numbers: “Pistachios | single | 1 | 0.8”). EFSA 2020 citation follows this value in text but governs “EFSA Nickel 2020 subsequently names nuts collectively…” — a distinct claim. |
| apple-cider-vinegar | Mitigation options | 5 ppb | ndungu2004-lead-vinegar-icpms-gfaas | not-in-source | Ndungu 2004 source page reports GFAAS MDL 0.12 µg/L for Pb, not ”≤ 5 ppb.” The ACV page cites Ndungu 2004 as “a useful method reference” for the protocol, not as the source of these detection-floor values. |
| apple-cider-vinegar | Mitigation options | 1 ppb | ndungu2004-lead-vinegar-icpms-gfaas | not-in-source | Ndungu 2004 is Pb-only; no Cd data or Cd detection limit in source page. |
| apple-cider-vinegar | Mitigation options | 10 ppb | ndungu2004-lead-vinegar-icpms-gfaas | not-in-source | Ndungu 2004 is Pb-only; no Ni data or Ni detection limit in source page. |
| apple-juice | Why this commodity accumulates heavy metals | 200 mg/kg | benoy1971-tin-toxicity-canned-fruit-juices | not-in-source | ”200 mg/kg” not stated in Benoy 1971 source page. Source page reports: Kuwait outbreak juice at 250–385 ppm (no symptoms in cats); controlled symptom onset at 540 ppm (1 cat), 1,370 ppm (20–30% cats; all humans). 200 mg/kg is below the no-symptom outbreak level. |
| apple-juice | Ranges by source, region, and variety | 1.5 ppb | fda2022-tds-elements-fy2018-fy2020 | not-in-source | Apple juice Pb p50 = 1.5 ppb (TDS Food 99, n=3). FDA TDS source page body does not enumerate per-food values; value exists in data/evidence/fda_tds_fy2018_2020_summary_by_food_analyte.csv. Ingredient page TDS Occurrence table confirms the value. |
| apple-juice | Ranges by source, region, and variety | 2.7 ppb | fda2022-tds-elements-fy2018-fy2020 | not-in-source | Apple juice Pb max = 2.7 ppb. Same structural gap as above; value in referenced CSV. |
| apple-juice | Ranges by source, region, and variety | 2.4 ppb | fda2022-tds-elements-fy2018-fy2020 | not-in-source | Apple juice tAs p50 = 2.4 ppb. Same structural gap; value in referenced CSV. |
| apple-juice | Ranges by source, region, and variety | 4.4 ppb | fda2022-tds-elements-fy2018-fy2020 | not-in-source | Apple juice tAs max = 4.4 ppb. Same structural gap; value in referenced CSV. |
| apple-juice | Ranges by source, region, and variety | 20 ppb | fda2022-tds-elements-fy2018-fy2020 | not-in-source | Apple juice Ni max = 20 ppb. Same structural gap; value in referenced CSV. |
| apple | Why this commodity accumulates heavy metals | 5.36 ppb | fda2022-tds-elements-fy2018-fy2020 | not-in-source | Apple tAs p90 = 5.36 ppb (TDS Food 78, n=27). FDA TDS source page body does not list this value; it exists in the referenced CSV and in the ingredient page’s inline TDS Occurrence table. |
| apple | Why this commodity accumulates heavy metals | 20 ppb | fda2022-tds-elements-fy2018-fy2020 | not-in-source | Apple tAs max = 20 ppb. Same structural gap; value in referenced CSV. |
| apple | Ranges by source, region, and variety | 5.36 ppb | fda2022-tds-elements-fy2018-fy2020 | not-in-source | Duplicate of Why-section claim above (same value, different section). Same structural gap. |
| apple | Ranges by source, region, and variety | 20 ppb | fda2022-tds-elements-fy2018-fy2020 | not-in-source | Duplicate of Why-section claim above (same value, different section). Same structural gap. |