Our methodology
The Margin Method
We don’t believe there’s one “best” homeschool curriculum. There’s the best one for your child, your prep capacity, your worldview, and your budget. The Margin Method is how we rank curricula by situation — and show our evidence at every step.
The promise
No curriculum can pay to rank higher. We take no payment, sponsorship, or placement fee from any publisher in exchange for ranking or position. Our scoring engine has no access to affiliate data — it literally cannot see who we earn from.
How rankings work
Every ranking answers one question: best for whom, under what conditions, and based on what evidence? For a given situation — a subject, a grade, a learning need, a worldview — we score each eligible curriculum on the dimensions that matter and rank by overall fit. We never crown a universal winner.
What data we use
- Curriculum & publisher data — grades, subjects, format, price, worldview, philosophy, prep level, independence.
- Structured fit attributes — how we classify each curriculum by learner type, rigor, prep, and special-needs suitability.
- Public parent sentiment — recurring themes from public discussions, summarized in our own words and attributed to their sources (never copied).
- Verified ProjectHomeschool reviews — first-party reviews from real parents, which carry more and more weight as they accumulate.
- Value analysis — what you actually get for the price.
How we use public parent sentiment
We read public parent discussions to spot recurring strengths and complaints — “great open-and-go option,” “heavy parent prep,” “strong phonics sequence.” This is deliberately careful:
- We only review a small allowlist of vetted sources, and we respect each site’s
robots.txt. We don’t crawl the open web. - We store a private snapshot for audit only — we never display copied review text, and raw third-party content is never published.
- What we keep is a summarized theme in our own words, attributed to its source.
- A human reviews and approves every theme before it can affect a ranking. Only approved sentiment counts.
- We call this “parent sentiment we reviewed” — never “verified reviews.” First-party ProjectHomeschool reviews are kept entirely separate.
Sentiment is one input among several, always paired with a confidence label so thin signals can’t masquerade as strong evidence, and refreshed as we re-review sources.
How we use ProjectHomeschool reviews
First-party reviews are the heart of the method. We only count reviews that are approved and verified. Until a curriculum has at least 5 verified reviews, we show “Not enough ProjectHomeschool reviews yet” rather than a guessed score — and dimension averages are computed only from those verified reviews.
Confidence labels — never fake precision
Every score carries a plain-language confidence label:
Research coverage: Strong — rich data and a healthy body of verified reviews.
Research coverage: Moderate — solid data and a growing set of reviews.
Research coverage: Early — early signal; treat with appropriate caution.
Not enough data yet — we won’t pretend to know yet.
Scores render as the Margin Mark — a single honey-intensity scale, never red/green grades:
Trade-offs always get their own line — they never quietly drag the score up or down.
How affiliate links work
Some outbound links are affiliate links, which means we may earn a small commission if you buy. This never affects scoring or ranking — affiliate relationships are excluded from the method by design, and disclosed wherever they appear. More on how we make money →
How often we update
Rankings are recomputed as new data arrives — new verified reviews, refreshed sentiment, and price changes — and each ranking page shows a real “last updated” date. We’d rather be honest about a limitation than overstate our confidence.