Method & sources

How we work

This is a record about real people, so it is built like evidence. Every published claim traces to a source; where the record is silent or contradicts itself, we say so rather than guess.

The order of evidence

When sources disagree, we show the disagreement and weigh it by this hierarchy, strongest first.

1. 1
   Primary, contemporaryThe gravestone itself; town vital records; church, probate, deed, and military records; original letters and photographs.
2. 2
   Contemporary compiledTown and church registers kept near the time of the events.
3. 3
   Later scholarly compilationThe 1893 History of the Town of Hingham genealogies, authoritative, but secondary, and itself sometimes copying earlier sources.
4. 4
   Modern aggregationsFind a Grave, BillionGraves, community trees, treated as leads to verify, not as proof.

How sure we are, claim by claim

Confidence attaches to the specific fact, not to a page-wide disclaimer. Confirmed means a strong source, or a match corroborated by an independent fact (a relative's name and an exact date both agreeing). Probable is shown hedged, "likely the same person." Possible matches are kept as open questions, not woven in as fact. Names repeat heavily here (214 Cushings, 75 Towers), so two same-named people are never merged just because a genealogy seems to fit.

Where the data comes from

The burial index was first gathered from Find a Grave and a public gravestone transcription, then matched against the public-domain 1893 History of Hingham genealogies. We treat the aggregators as a starting point and move toward primary sources as each life is verified. We don't claim two datasets are "independent" unless we can show it, and we link to grave photographs on their source rather than re-hosting work that belongs to the people who took it.

When a source is wrong, we say so, and fix it

Find a Grave is a lead, not the last word. Its memorial for Capt. Theophilus Cushing (1657–1718) titled him "Rev.", but the 1893 History of Hingham and the 1905 Cushing Genealogy both record him as a farmer and selectman; the ministry belonged to his brother, Rev. Jeremiah Cushing. We corrected the record, kept the reason and the sources beside it, and let the fix carry through to every relative who links to him. Corrections are treated as evidence, not quiet edits.

Where a person appears in the 1893 genealogies, we now read their own family entry and set out what it says, parents, marriage, trade, children, each fact only as sure as the match itself, and each shown with how the match was made (an exact birth date in their own numbered entry, corroborated by a named relative). A common name matched on weaker evidence is marked an open question, never woven in as fact.

What AI does, and doesn't

New tools help us locate, transcribe, correlate, and draft, and a person checks every published claim against a cited source. They are never used to invent a fact, generate a face or voice, or put words in the mouths of the dead. The charter holds us to that →

Found an error, or hold a source we don't? That's the work, not an interruption. Here's how to help →