A Family History Research Log That Tracks What's Proven and What's Remembered

Most family history research ends up split across two places that never talk to each other: a genealogy platform full of certificates and census rows, and a separate pile of notes from phone calls with relatives who remember things no document ever recorded. The two rarely get compared, so contradictions sit unnoticed for years, and a good story from Aunt Carol has no fixed place to live once the call that produced it is over.

A family history research log is a single running record that holds both kinds of evidence — documented and remembered — side by side, with enough detail to tell them apart. It tracks what each claim is, where it came from, how confident anyone is in it, and whether it agrees or conflicts with anything else already logged. The point isn't to settle every disagreement. It's to stop losing track of who said what, and to keep an unresolved contradiction visible instead of quietly picking whichever version sounds more certain.

Why a spreadsheet of dates isn't enough

Genealogy software is built to hold documentary evidence well: a birth certificate, a census entry, an immigration record, each with a source citation and a date. What it isn't built to hold is the other half of the research — the phone call where a cousin says the family didn't actually arrive the year the ship manifest suggests, because their grandmother always insisted they'd been in the country two winters already when that boat came in.

That kind of claim doesn't fit a source-citation field designed for a document, so it either gets left out or gets written down somewhere that isn't connected to the record it contradicts. Either way, the contradiction disappears instead of getting resolved or at least remembered as open. A research log fixes this by giving oral claims the same structural treatment as documentary ones — a place, a source, a date, a confidence level — rather than treating them as anecdote that lives outside the real research.

Two kinds of evidence, not one ranked above the other

Genealogists already have language for this distinction. The Genealogical Proof Standard treats documentary sources and oral tradition as two different categories of evidence, not a hierarchy where enough of one eventually outweighs the other. A document can be wrong — clerks misheard names, ages got rounded, immigration dates got recorded on arrival rather than on the crossing itself. A memory can also be wrong, reshaped by decades of retelling. Neither one is automatically the tiebreaker.

A research log should record both without silently privileging the document just because it has a government seal on it. What it should record honestly is the kind of evidence each claim is, so that whoever reads the log later — including the person who wrote it — can weigh it accordingly instead of mistaking a firm family story for a verified fact, or a clerical record for the full truth.

What to track for every claim

Each entry in the log is one claim about one person, place, or event — not one person's entire biography. A single ancestor will usually have several entries: one for a birth date, one for why the family left a particular town, one for a name that shows up spelled two different ways across two documents. The fields that make an entry usable months or years later:

Field	What it captures
Claim	The specific fact or story, stated plainly
Person / event	Who or what it's about
Source type	Documentary record, told directly by someone with firsthand knowledge, or told secondhand
Source detail	The record's name and location, or who said it and when
Date recorded	When this entry was logged, not when the event happened
Confidence	Certain · fairly confident · best guess · unverified
Agrees or conflicts with	Any other logged entry this supports or contradicts
Status	Resolved, open, or not worth resolving

The "agrees or conflicts with" field is the one most research logs skip, and it's the one that turns a list of facts into something that actually catches contradictions instead of just filing them next to each other.

Recording a disagreement instead of resolving it

Some contradictions can be resolved — a second document turns up, or someone realizes a birth year was rounded for a form. Many can't be, and a research log needs a way to hold those permanently rather than forcing a decision that isn't actually available. If one aunt says the family left in 1952 and a cousin says 1954, and no document settles it, the honest entry records both accounts, who said each one, and marks the status as open — not a coin flip dressed up as certainty.

This matters for the same reason it matters in any interview: a remembered account is not the same kind of evidence as a dated record, and two people can disagree about a decades-old year without either one being wrong. The Oral History Association's published guidance on narrator review — reading an account back to the person who gave it, and letting them correct or add to it before treating it as final — is worth borrowing here even outside a formal interview. A quick "so you're saying it was before the move, not after?" during the call itself, logged alongside the answer, does most of the work of catching a slip before it hardens into a fact nobody questions again.

Using the log during an interview, not just after

A research log works best brought into the conversation, not filled in from memory afterward. Before a call with a relative, scan the log for open items connected to them — the contradiction nobody's resolved, the claim with only one source, the name spelled two ways. Ask about those specifically instead of starting from a blank page. Most relatives can't produce useful detail from "tell me what you know about the family," but they can usually react directly to "your cousin says we left in '52 — does that sound right to you?"

After the call, log the new entry the same day, while the exact wording is still available. A story compressed into a summary a week later loses the qualifiers — the "I think," the "unless I'm mixing it up with" — that are often the most honest part of what someone said.

A log entry to copy

For each claim, record:

• Claim: the specific fact, stated in one sentence
• Person or event: who or what it concerns
• Source type: documentary record / told firsthand / told secondhand
• Source detail: record name and location, or who said it and when
• Date recorded: today's date, not the event's date
• Confidence: certain, fairly confident, best guess, or unverified
• Agrees or conflicts with: any related entry, by name
• Status: resolved, open, or not worth resolving
• Anything the source asked not to be shared further

That last line belongs in the template, not as an afterthought. A relative who says "you can write that down, just don't put it in anything you show my sister" has given you information with a condition attached, and the log should hold the condition as carefully as it holds the claim.

Where this fits inside the larger archive

A research log is a working document, not a finished one — it stays open as long as the research does, which for most families is indefinitely. The archive is the product, not the book makes the broader case for treating the ongoing record as the thing worth building, with any printed or bound version as one export of it made later. What to do when the family tree has names but no people covers the companion problem — a tree can hold the dates a log verifies, but it takes a deliberate pass through the people still living to attach a story to each name before the log's open items become permanently unresolved.

Joan works from the same distinction a good research log keeps: a remembered account and a documented fact are both worth keeping, logged as what they actually are, never smoothed into a single certain version.

If your family is ready for someone else to hold that pacing and the follow-up, join Founding Families.