Methodology

Dataset integrity

The active site file contains 1588 unique observations across 124 institutions, retrieved October 5, 2026. It is the national merged dataset; the previous site dataset is not loaded.

Inventory counts include every source row before state, confidence, freshness and numeric-value filters. Default comparison tables show only eligible, rankable observations.

Collection

LoanMate records publicly advertised figures from lender-owned rate pages and preserves the direct source, verification date and any stated assumptions. The accompanying directory contains 88 active Massachusetts lenders plus national sources represented by the merged dataset.

Plausibility verification gates

A collected figure must pass plausibility gates for its product before it can appear in a ranking: new-vehicle loan APRs between 4.00% and 10.00%, used-vehicle between 4.00% and 12.00%, unsecured personal loans between 5.90% and 20.00%, and secured personal loans between 3.00% and 15.00%. A figure outside its band is held for re-verification against the lender's page instead of being published as-is; verified exceptions are documented with their assumptions.

Categories and state eligibility

Home, auto and personal loans are separated before ranking. Home Loans then separate purchase mortgages, refinances, HELOCs and fixed home equity loans. Auto Loans separate new and used vehicles. Personal Loans separate secured and unsecured products. The visible borrower-state selector removes Massachusetts-local institutions outside Massachusetts and applies verified product-specific exclusions; an unresolved authorization status is a reason to omit, not assume eligibility.

Ranking

Organic tables respond only to the visible state, product, term, institution and sort controls. Compensation never changes organic order. Home-loan tables can sort by rate, annual percentage rate or zero points; auto and personal tables sort by advertised APR.

Verification dating on every rate table

Every comparison table displays its dataset refresh date, and every rate row carries its own last-verified timestamp alongside the lender source. A source more than seven days old stays in the inventory for auditability but is excluded from default rankings and labeled as stale where it appears.

Confidence and freshness

High-confidence rows have a clear lender source and product mapping. Medium-confidence rows retain their caveats. Low-confidence rows are excluded from default rankings and placed in a less-verified section. A source more than seven days old is retained in the source inventory for auditability but is not rendered in comparison views.

Disclosure treatment

Payment amounts are intentionally omitted from comparison tables. Rate and annual percentage rate are shown together when both are available, and ARM rows state that the rate may increase after the fixed period. HELOC rows pair the advertised rate with available fees, variable-rate status and maximum APR; missing source terms are labeled as not listed.

Refinance labeling

If a lender’s purchase table is used as a refinance comparison assumption, the row is labeled purchase-table assumption — not genuine refinance pricing. It is not represented as a lender-quoted refinance offer.

Calculators

The seven calculators use only numbers entered on the page. Their results are educational estimates, not offers, approvals, guarantees or lender quotes. No calculator result is pre-populated from a listed lender.

Corrections process and retention

If a rate appears wrong, we recheck the lender's source, product mapping, eligibility notes, and verification date. While a material issue is unresolved, the observation is removed from default rankings. A corrected row receives its actual re-verification date. Dataset versions and dated observations are retained so published figures can be reconstructed. Errors can be reported using the contact details on the Contact page, whose team reviews corrections under the Editorial & corrections policy.