Every Canadian competitor leans on a single rating. We combine four independent layers, publish the math, and show you what each one contributes. If you don't like our methodology, you should know exactly why.
Each layer's score is normalized to 0–100, then weighted and summed. The Roost field-research layer — our manually-curated signals from news, school council posts, and community sources — is applied as a small adjustment (typically ±0 to ±5), capped at 95 to prevent any single field-research flag from washing out the underlying signal.
When a layer is missing (e.g., a school with no AEAM survey data yet), the formula redistributes the missing weight pro-rata across the layers we do have — and we clearly tell you which layers contributed (look for the layer indicators on each school card).
Tiers (S/A/B/C/D) are cut points within the composite scale — calibrated as a continuum, not as pass/fail. A "C" school is mid-pack, not failing.
Province-wide-standardized signal. 3-year smoothing controls for cohort noise. Refreshed annually.
Susceptible to exemption-gaming. Misses non-tested grade-bands (K, 1, 2, 5). Doesn't measure citizenship, well-being, or specialty programs.
Commonly cited; familiar to parents. Comparable across years. Authoritative source.
Correlates ~0.56 with neighbourhood SES (Raptis 2012, Clothier 2011). Low-SES catchments cannot easily out-rank high-SES even with strong instruction — Fraser-alone misleads.
Parent-reported lived experience. Surfaces Concerns + Issues that test scores miss. Province-wide instrument, annually administered. No competitor uses this.
Self-reported; subject to recall bias. Small samples at small schools. SIRR PDFs require parsing (which is why nobody else surfaces it).
Catches the stuff a PDF can't: incidents, leadership churn, grade truncation, capacity pressure. Quarterly refresh from news + Reddit + school-council posts.
Manually curated → some subjectivity. Lags by up to 3 months. Capped at ±5 to prevent any single flag from dominating.
The full published methodology — what each layer measures, how we weight it, and which sources we deliberately distrust.
Fraser, EQAO, and Local Logic each have well-documented blind spots (SES confounding, churn, sponsor bias). Naming those blind spots in writing is the price of asking parents to pay for a decision.
We're independent — no school board, realtor, or developer pays us. We publish the composite formula, the layer weights, and what we explicitly exclude. Read it once, audit it forever.
A tool you can audit, not a black box. If you find a flaw, you can name the line that's wrong — and we'll fix it.