We checked 33 holiday plant articles against ASPCA records. Two sentences disagreed.

Published October 13, 2026 · By Zhe Li, editor

Across 43 sampled articles, 33 accessible articles yielded 472 statement units: 3 candidate units differed from the ASPCA record, 109 were consistent, and 360 were excluded under the recorded rules.

What we checked

The sample came from a mechanical search-query rule registered before any sampled article was read. Queries had to name a holiday plant, a toxicity or safety term, and cats or dogs. They were grouped by plant and animal set, ranked by US monthly search volume, with alphabetical tie-breaking. The highest-volume query from each of the top 10 groups was used. There was no animal quota. The keyword file had been collected for cat-side demand, and its volumes were annual monthly averages, rather than December figures.

For each query, the sample took up to 10 organic results from an Ahrefs search snapshot and deduplicated them by URL. The resulting 73 URLs comprised 43 checkable articles, 23 user-generated or video results, and 7 reference results. The latter classes remained listed but were not checked. Inaccessible articles also remained in the sample.

We compared explicit plant-and-animal statements with ASPCA plant entries, reading every toxicity or non-toxicity line for the relevant animal. Pet Poison Helpline was checked for conflicting animal-specific records where archived material was available. Verdapaw's S0–S4 grades rest on ASPCA and Pet Poison Helpline records; this comparison concerns toxic/non-toxic direction, with disagreements about degree excluded.

What we found

The 3 candidates represent 2 distinct article sentences. The following table preserves the wording, including the article's spelling and punctuation. All quoted article and reference snapshots below were accessed on 2026-10-08 UTC.

For Article 12, the ASPCA entry names Euphorbia pulcherrima. Pet Poison Helpline's archived poinsettia entry reads: “Though they have a bad rap, poinsettia (Euphorbia pulcherrima) plants are only mildly toxic to cats and dogs.” Both references name cats and dogs and give a toxic direction; the article sentence gives a non-toxic direction.

The same article section also says: “The white sap that excretes when a leaf brakes off is the only part to be mildly toxic to both cats and dogs.” We retain both sentences. The section contains contradictory statements; one differs from the ASPCA record. We do not assign an overall classification to the article.

Article 12 evidence: article URL. References: ASPCA entry, Pet Poison Helpline entry.

For Article 14, the article and ASPCA entry both name Schlumbergera bridgesii. The article's “technically poisonous to cats” differs from the ASPCA's cat-specific non-toxic record. An archived Pet Poison Helpline paragraph says: “Even though this plant is considered to be non-toxic to pets, ingestion can cause mild GI distress (vomiting, diarrhea).” That paragraph says “pets”, rather than naming cats or dogs, so it does not establish a cat-specific binary direction. It is quoted as context, not as an animal-specific conflicting record.

Article 14 evidence: article URL. References: ASPCA entry, Pet Poison Helpline paragraph.

Archived copies of every quoted page, accessed 2026-10-08 UTC, are held by Verdapaw and available on request.

How sure we are

The strict count is 3 candidate units; the recorded sensitivity range is 3–9. The upper bound adds only the adjudicated alternatives: B115 adds 1 unit, A080 adds 2, A083 adds 1, and original row 97 adds 2. All remain excluded in the final classifications. This upper bound describes alternative readings rather than additional accepted findings.

Original rows 114–116, concerning “low in toxicity for cats”, are degree-only exclusions and are not added to the sensitivity upper bound.

The earlier blind comparison recorded agreement on 148/159 units, or 93.1%. The later independent review recorded 151/159, or 94.97%, which is also the final agreement figure. It is below the recorded 95% threshold. The remaining disagreements followed the specified editorial-adjudication route. There were 10 editorially adjudicated units: 8 in the blind sample and 2 from newly accessible articles.

Limits and disclosures

There were 10 inaccessible articles. Under the previously recorded retry procedure, 5 articles initially returning HTTP 429 received a single retry after at least 24 hours. Of these, 4 became accessible and 1 still returned 403. The newly accessible articles received a separate full blind review after the original blind review.

Amendments 1–5 were written before any sampled article was read. Amendment 6 was written after the first pass, to resolve 143 statements the pre-registered rules did not cover. It reclassified them as consistent or excluded and added no discrepancies. Amendment 7 was written after the first pass and the second-pass blind review.

Independence between review passes was limited; details are in the methods appendix.

This report compares archived wording under the stated rules and makes no toxicology claims beyond the reference records.

The author is an editor, not a veterinarian. Verdapaw had no advertising when this report was published. If your pet ate a plant, contact your vet or a poison hotline.

Method link

The public methods record, including the dated amendments and withdrawn provision, will be linked at /reports/holiday-plants-2026/methods.