How we rank
How Project Homeschool ranks curriculum
There’s no single “best” homeschool curriculum — only the best one for your child, your teaching style, your worldview, and your budget. We compare curriculum by situation, on three separate kinds of evidence, and we show our confidence at every step.
Researched and maintained by the founders of Project Homeschool — homeschool parents building this in the open. About us →
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.
Three kinds of information, kept separate
- Curriculum facts — grade range, subject, cost, format, worldview, parent prep, and student independence, verified by a human.
- Public parent sentiment — recurring themes from publicly available reviews, discussions, videos, and articles, summarized in our own words and attributed to their sources (never copied).
- Verified parent reviews — first-party reviews submitted directly to Project Homeschool, which carry more and more weight as they accumulate.
When verified reviews are limited, we label the page as an early signal and show data confidence clearly — instead of inventing a precise score we can’t back.
How rankings work — two stages
Every ranking answers one question: best for whom, under what conditions, and based on what evidence? We never crown a universal winner. Each ranking runs in two strict stages:
- Eligibility (a hard gate). Each situation declares who belongs. A math ranking only includes programs that actually teach math; a secular ranking excludes anything with faith integration; an all-in-one ranking excludes single-subject programs. A subject or worldview mismatch is a hard exclusion.
- Fit (a situation-weighted ranking). Only programs that pass the gate are scored, and the weighting changes by situation — a struggling-learner math page leans on low prep and gentleness, an independent-learner page leans on how well a child can work alone.
Because we don’t yet have a large body of verified parent reviews, we show an honest fit band— Strong, Good, Mixed, or Early signal — rather than a precise decimal we can’t back.
How we use public parent sentiment
We read public parent discussions, reviews, and videos to spot recurring strengths and complaints. 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. We paraphrase patterns in our own words, attributed to the source.
- A human reviews and approves every theme before it can affect a ranking. Only approved sentiment counts.
- It impacts a score only lightly and is always paired with a confidence label, so thin signal can’t masquerade as strong evidence.
How we use verified parent reviews
First-party reviews are the strongest signal we have. We only count reviews that are approved and verified. Until a curriculum has enough of them, we show “Verified parent score pending” rather than a guessed score.
Confidence labels — never fake precision
Every profile carries a plain-language data-confidence label:
Strong signal — curriculum data, meaningful public sentiment, and enough verified reviews for stronger confidence.
Developing signal — useful data and public sentiment, but verified reviews are still limited.
Early signal — limited data and/or limited public sentiment; treat as a starting point.
Scores render on 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 fit scores, rankings, review summaries, comparison results, or recommendations — affiliate relationships are excluded from the method by design. 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.