Genealogy forums talk about the Genealogical Proof Standard as if it’s a checklist. Five things. Tick them off, mark the fact “proven,” next.
I wish. The GPS isn’t a checklist, it’s a definition of “done.” And the definition is more demanding than it sounds.
- Reasonably exhaustive search
You have to look everywhere a competent genealogist would look. Not where Ancestry suggested. Not where a green leaf pointed. Everywhere — federal records, state records, county records, church registers, newspapers, the stuff that isn’t indexed, the stuff that doesn’t show up in a surname search.
“Reasonably exhaustive” is not “infinitely exhaustive. You stop eventually. But when you stop, you should know what you didn’t search and why — not “because I got bored.”
- Complete and accurate source citations
Someone else should be able to walk into the same repository and pull the same record using nothing but your citation. Repository name. Record group. Volume. Page. Image number. Date you accessed it.
A link to a search-results page isn’t a citation. “Ancestry.com” isn’t a citation. “FamilySearch” isn’t a citation.
- Analysis and correlation
You have to actually think about what each source says. Original or derivative? Did the informant witness the event or hear about it secondhand? Does this record answer the question directly, or only when you combine it with something else?
Then you lay the sources side by side. Do they agree? Fill each other’s gaps? Contradict each other?
Most people skip this. They find one record, take it at face value, and move on.
- Resolution of conflicting evidence
This is the one that kills trees. Two records disagree on a birth year: one says 1842, one says 1843. Easy to pick one and ignore the other. The GPS doesn’t let you. You have to write out why you chose one over the other. Was one informant closer to the event? Is one record a derivative of the other? Does a third source break the tie?
You don’t ignore the inconvenient record. You explain it.
- Sound written conclusion
Not a checkbox. Not a number on a confidence slider. A paragraph, maybe a page, that another researcher can read and evaluate for themselves. Here’s the answer, here’s the evidence, here’s where the evidence conflicted and how I resolved it. Judge the reasoning for yourself.
Why I care about this
None of the genealogy software I own does any of this as a workflow. RootsMagic, Family Tree Maker, Ancestry, MyHeritage — they’re all designed around entering conclusions. You type in a birth date, attach a source, maybe add a note, and you’re done. The reasoning doesn’t live in the software. It lives somewhere else — a notebook, a Word doc, a research log.
That’s why I keep my research log separate from my tree. RootsMagic holds the verdict. The log holds the trial.
What this means for the blog
My first post explained that I’m starting from zero: 1,800 people in my tree, almost nothing verified, first real step is my own birth certificate.
This post is the standard those 1,800 people are going to get measured against. Some posts here will be about process: how to actually search a probate file, what makes a citation usable, what to do when records disagree. Some will be about findings — a real ancestor, a real record, the evidence laid out so you can check it yourself.
A lot of the posts are going to sound cautious. They’re going to qualify things that other genealogy blogs state as fact. If you’re reading this and wondering why, this post is why. I’d rather be right than be fast.