The Genealogical Proof Standard: Sourcing Evidence in Family History
The Genealogical Proof Standard (GPS) is the framework that separates serious research from a collection of names. Here\'s the five-element standard, why each matters, and how to apply it.
What the GPS is
The Genealogical Proof Standard (GPS) is the methodology that separates serious genealogical research from a collection of plausible-looking names. It was codified by the Board for Certification of Genealogists and is the standard applied to credentialed genealogists, court genealogical evidence, and lineage-society applications.
The GPS has five elements. A conclusion meets the standard when *all five* are satisfied:
- Reasonably exhaustive research.
- Complete and accurate source citations.
- Analysis and correlation of the collected information.
- Resolution of any conflicting evidence.
- A soundly reasoned, coherently written conclusion.
The five together are what produce a genealogical *proof*, as opposed to a guess or a hint.
1. Reasonably exhaustive research
"Reasonably exhaustive" does not mean every record in existence - it means every record that a competent researcher would reasonably search before concluding. Typically that means: vital records (BMD), census records spanning the person's adult life, parish registers if applicable, probate records, immigration/naturalisation if relevant, and at least one independent source for every key fact.
A single census entry is *not* exhaustive research. Two census entries from different decades are weak. Two census entries plus a vital record and a probate document are getting close.
2. Complete and accurate citations
Every fact should have a citation specific enough that another researcher could locate the same record. That typically means: the archive or database name, the collection or record series, dates and reference numbers where applicable, and the URL or shelf mark.
"Family tradition" and "found on Ancestry" are not citations. "Ancestry.com, England & Wales, FreeBMD Marriage Index 1837-1915, vol. 1a, p. 234" is.
GEDminer's sourcing audit lists every fact in your tree without a citation, ranked by importance. See the citing genealogy sources guide for templates.
3. Analysis and correlation
Records have to be evaluated, not just collected. Two records can both be "true" yet support different conclusions if they are about different people with similar names. Correlation means lining up multiple records and asking: do they all support the same conclusion?
A classic correlation test: do the ages on three census records add up to a consistent birth year? If 1851 says age 20, 1861 says age 28, and 1871 says age 42, you almost certainly have two different people conflated.
4. Resolution of conflicting evidence
Conflicts are inevitable. The GPS does not require they never occur - it requires you address them. If two records give different mother's maiden names, do not silently pick one. Document the conflict, weigh the relative reliability of the two sources, and explain the chosen conclusion.
Unresolved conflicts left in a tree are landmines for future researchers (including you, six months from now).
5. A soundly reasoned, written conclusion
For each non-trivial conclusion, write down *why* you believe it. A short note attached to the individual record explaining "I conclude this John Smith is the same as the John Smith in the 1851 census because..." is enough.
This discipline forces you to confront whether your evidence actually supports the conclusion. Half the time, the act of writing it out reveals you had less proof than you thought.
Applying the GPS in practice
You will not apply the full GPS to every record - that would be impractical. Apply it rigorously to:
- The earliest ancestor on each direct line (these anchor everything below them).
- Anyone you intend to publish or share.
- Anyone with conflicting evidence in the tree.
- Lineage-society applications.
For everyday research, aim for at least one solid source on every vital fact and a written note for every non-obvious conclusion. The GPS sits on top of that as the standard you reach for when something matters.
How GEDminer helps
GEDminer's sourcing sub-score (30% of the Tree Health Score) measures (Total Sources / Total Facts) × 100, capped at 100 - a direct numerical proxy for how well-sourced your tree is. The audit lists every unsourced fact, ranked by importance, so you can prioritise the GPS-relevant ones first.
See the data quality score guide for how the score is calculated, and the common genealogy mistakes guide for the GPS-related pitfalls most often seen in real trees.