Methodology

How plant data is curated, how citations work, what "honest" means in our context, and how we handle disagreements between sources.

This page documents how content on Outdoor Plant Care is created, fact-checked, and updated. Read it once and you'll know exactly how to evaluate what we publish here.

The citation hierarchy

Every factual claim on this site is cited against a primary source. When sources disagree, we use this hierarchy in order:

1. Federal and university research publications \u2014 USDA, NOAA, university Extension publications. Highest weight.
2. Major botanical garden plant databases \u2014 Missouri Botanical Garden PlantFinder, NC State Extension Plant Toolbox, Royal Horticultural Society, Chicago Botanic Garden. These curate from primary research with editorial oversight.
3. Peer-reviewed horticulture journals \u2014 HortScience, HortTechnology, journals of state horticultural societies. Used for specialized claims.
4. Specialty programs \u2014 Texas A&M AgriLife Earth-Kind for roses, Xerces Society for pollinators, USDA NRCS for soil, USGA for turfgrass. Used in their domain.
5. Thomas's firsthand garden observation \u2014 only for the plants he grows in zone 7a Long Island, only for first-person passages. Never overrides 1-4 for universal claims.

We do not cite Wikipedia, gardening blogs, nursery marketing copy, or social media. Those may be correct, but they are not primary.

How a plant guide is built

For a new species guide, here is the actual workflow:

1. Compile primary references. Pull the species entry from Missouri Botanical Garden, NC State Plant Toolbox, Penn State Extension, and any other Extension publication that covers it.
2. Identify disagreements. Note where sources give different zone ranges, sun requirements, mature size, or care timing. There are always some.
3. Resolve via the hierarchy. Where federal/Extension sources disagree, prefer the source closest to the geography the user is reading from. For zone ranges, take the broadest cited range and note edge cases.
4. Write from sources. Every paragraph that makes a factual claim includes a `Per Source Name` citation. We do not paraphrase without crediting.
5. Add firsthand context. Where Thomas grows the plant in zone 7a Long Island, first-person observations are added, clearly distinguished from sourced material.
6. Quality check. Banned-word scan, exclamation-point check, citation density check (typically 5-15 inline citations per 2,000-word guide).
7. Visual audit. Hero image checked for accuracy (a "best peony cultivars" article gets a peony photo, not a rose).

What "honest" means here

We use "honest" as a deliberate label, not a marketing claim. Specifically:

• No fabricated personal experience. First-person voice is only used for plants Thomas actually grows. The list is on the about page and audited.
• No fabricated product testing. Buyer guides recommend products either because Thomas has used them or because they are the universal first recommendation from Extension publications. We label which is which inline.
• No AI-generated filler. Every line is hand-written by the editor or hand-curated from cited sources. Articles are not generated by language models and then "edited." That route was explicitly rejected.
• No paid placements. No company has paid for inclusion. If that ever changes, sponsored material will be labeled at the top of the article.
• No "in conclusion" wrap-ups. No "leverage", "embark on a journey", "delve into", or any of the other AI-marker phrases. The full banned list is in our internal style guide.

Handling disagreements between sources

The most common sources of disagreement:

• USDA zone ranges. Catalog descriptions over-promise. Extension publications often give narrower, more honest ranges. We prefer Extension.
• Mature plant size. Nursery tags routinely understate this by 20-50%. We use Missouri Botanical Garden or NC State numbers, which reflect landscape-scale reality.
• Pruning timing. Genuine regional differences exist. We give the rule and note exceptions by zone.
• Pest treatment effectiveness. Marketing claims of biological controls vary widely. We default to the verified protocols from UC IPM, Cornell's NYSIPM Program, or NCSU's pest fact sheets.

When the disagreement is genuine and unresolvable, we say so. "There is debate about whether basil actually repels hornworms, but it doesn't hurt anything."

Updates and corrections

When science changes (e.g., the 2023 USDA Plant Hardiness Zone Map updated from 2012), we update affected articles and log the change on the changelog.

When a reader points out an error, the correction is made and acknowledged inline at the bottom of the corrected article. We don't memory-hole mistakes.

How to verify what we say

Every cited source on this site is linked. If you want to verify a claim, click through to the cited source and read the original. The sources page lists the full bibliography of references used across the site.

If you find an error, tell us and we'll fix it.