Every Family Group Sheet in the DFA’s database rests on an unglamorous foundation: a citation. Behind “Johann Heinrich Dillmann, born circa 1830, Hesse, Germany” sits a church record, a passenger manifest, a census taker’s handwriting, or sometimes only a family story typed up by a great-grandchild decades after the fact — and those are not the same kind of thing. One of the quiet purposes of this series has been to show what a documented family history looks like when the underlying sources are treated with real rigor. This article steps back to explain where that rigor comes from: the field of genealogical evidence analysis, and the book most responsible for codifying it, Elizabeth Shown Mills’ Evidence Explained.

Why All Sources Are Not Created Equal

Evidence Explained opens with a question genealogists rarely stop to ask themselves: why cite sources at all? The book’s answer is blunt — because no historical record can be trusted at face value, and the only way to judge how much to trust it is to know exactly what it is.^[1] A birth certificate filled out by an attending physician the day of the birth is not the same kind of evidence as a family Bible entry written years later from memory, even if both say the same thing. Mills built her framework around three separate, overlapping classifications that every genealogist learns to ask of every record:

Source classification asks whether a record is original (the first recording of information, such as the actual parish register page) or derivative (a copy, transcription, abstract, or database entry built from an original). A transcription can introduce errors that don’t exist in the source it was copied from — a common concern with the online indexes most family historians rely on daily.

Information classification asks whether the information within a record is primary (supplied by someone with firsthand knowledge, recorded close to the time of the event) or secondary (supplied by someone without firsthand knowledge, or recorded long afterward). A death certificate typically contains both: primary information about the death itself, reported by an attending physician, and secondary information about the deceased’s birth date and parents’ names, often supplied by a grieving relative working from memory.

Evidence classification asks whether a piece of information directly answers the research question at hand (direct evidence), only implies an answer that must be reasoned out from context (indirect evidence), or is the meaningful absence of an expected record (negative evidence — a family that should appear in a census and doesn’t, for instance).

Citing a worthless source is an effort that produces worthless results. The validity of any piece of evidence cannot be analyzed if its source is unknown.^[2]

None of these three classifications are about whether a source is “good” or “bad.” A derivative source with secondary, indirect information can still be useful — it just has to be weighed accordingly, and weighed against everything else available, rather than accepted as settled fact because it appeared in a database with a confident-looking search result.

The Genealogical Proof Standard

Sitting alongside Evidence Explained, and closely associated with Mills’ teaching, is the Genealogical Proof Standard (GPS) — the Board for Certification of Genealogists’ five-part definition of what it takes for a genealogical conclusion to be considered credible.^[3] The five elements are:

1. Reasonably exhaustive research. Not every record that could conceivably exist, but enough of the relevant records that an important piece of contradicting evidence is unlikely to have been overlooked.
2. Complete and accurate source citations for every fact, so the research can be verified, checked, and built upon by someone else.
3. Analysis and correlation of the assembled evidence — not just collecting records, but reading them against each other.
4. Resolution of any conflicting evidence, rather than quietly picking the version of events that’s most convenient.
5. A soundly reasoned, coherently written conclusion, understood to be provisional and open to revision if new evidence emerges.

That last element matters more than it might seem. A GPS-compliant conclusion is never claimed to be permanent. It is the best answer the current evidence supports — which is precisely the spirit behind the “flag conflicts, don’t silently resolve them” approach this site has taken toward its own Family Group Sheets data, including the still-open numbering conflict at Groups 164–167 and the incomplete records at Groups 46, 105, and 129.

A Brief History of Evidence Explained

Elizabeth Shown Mills has spent decades teaching genealogical research methodology, launching the Advanced Research Methodology track at Samford University’s Institute of Genealogy and Historical Research in 1986 — a course structure that became the model many genealogical education programs still use.^[4] Her first citation guide, the considerably shorter Evidence! Citation & Analysis for the Family Historian, appeared in 1997. It sold well enough, and generated enough follow-up questions from readers encountering record types the book hadn’t anticipated, that Mills spent years afterward answering one-off citation questions by email before finally sitting down in 2004 to build something comprehensive. The result, published in 2007, was Evidence Explained: Citing History Sources from Artifacts to Cyberspace — now in its fourth edition (2024), and containing well over a thousand citation templates covering everything from medieval manuscripts to social media posts.^[5]

The book’s templates are built on the Chicago Manual of Style’s citation logic, extended to cover the specific, often idiosyncratic record types genealogists actually use — church registers, passenger manifests, probate files, DNA test results, and digitized archives that standard style guides simply don’t address.

Why the DFA Cites the Way It Does

None of this is academic throat-clearing for a project like this one. The DFA’s Family Group Sheets span more than two centuries, a dozen countries, and at least five distinct spelling traditions of a single surname — exactly the kind of dataset where sloppy sourcing compounds into real errors. A transcription error in one Family Group Sheet, cited confidently enough, can propagate into every article and every researcher’s notes that build on it afterward. The reverse is also true: a well-cited fact, even an inconvenient or incomplete one, remains useful precisely because the next researcher can trace it back to its source and judge it for themselves.

This is the same reasoning behind conventions readers of this series have already seen in practice: transcribing source typos verbatim rather than silently correcting them, flagging data conflicts (like the 164–167 numbering collision) for human resolution instead of guessing at a resolution, and treating every statistical claim in the “What the Numbers Tell Us” series as provisional, revised openly when a larger dataset or a closer reading changes the picture. That is the Genealogical Proof Standard’s fifth element in practice: a conclusion is only as good as the evidence behind it, and it stays open to correction.

For Dillman Researchers

A few practical habits, drawn directly from Evidence Explained and the GPS, are worth adopting for anyone researching their own Dillman line:

• Ask what kind of source you’re looking at. An index entry on a genealogy website is a derivative source built from an original record you haven’t seen yet — treat it as a clue pointing toward that original, not as the original itself.
• Separate the primary from the secondary information within a single record. A death certificate is strong evidence of a death date and place; it is much weaker evidence of a birth date reported decades earlier by a surviving spouse.
• Write down where you found everything, immediately, in enough detail that you could find it again in five years. This is the single habit genealogists most often regret not building early.
• When two sources disagree, don’t just pick the one you like better. Note the conflict, and see whether a third source resolves it. If nothing resolves it, say so in your own notes rather than silently choosing one version.
• Treat every family group sheet — including the DFA’s own — as a working hypothesis, not a finished answer. That’s exactly how this site tries to treat them too.

Family history research doesn’t require a certification to do well. It requires the willingness to ask, of every fact, where it came from and how sure anyone can really be. That question is the entire discipline in miniature — and it’s the same question this site tries to ask of its own 199 family groups, article after article.

Sources

1. Elizabeth Shown Mills, Evidence Explained: Citing History Sources from Artifacts to Cyberspace, 4th ed. (Baltimore: Genealogical Publishing Company, 2024), chaps. 1–2.
2. Elizabeth Shown Mills, Evidence! Citation & Analysis for the Family Historian (Baltimore: Genealogical Publishing Company, 1997), as summarized in publisher and bookseller descriptions; see masthof.com (accessed July 2026).
3. Board for Certification of Genealogists, Genealogy Standards, 2nd ed. rev. (Nashville: Ancestry, 2021), 1–8; summarized at bcgcertification.org/ethics-standards (accessed July 2026).
4. “Elizabeth Shown Mills,” FamilySearch Research Wiki (accessed July 2026).
5. Diane Haddad, “How and Why to Cite Sources: An Interview with Elizabeth Shown Mills, Citation Guru,” Family Tree Magazine (accessed July 2026); “Evidence Explained, 4th Edition,” Genealogical.com (accessed July 2026).

AI Disclosure: This article was researched and written with the assistance of Claude, an AI assistant developed by Anthropic. Historical and biographical claims about Elizabeth Shown Mills and Evidence Explained are drawn from the sources cited above; DFA-specific claims are drawn from the association’s own published conventions and prior articles in this series.

Notes

1. References to the DFA’s Family Group Sheets database and prior articles in this series (the “What the Numbers Tell Us” series and its treatment of the Groups 164–167 numbering conflict) are drawn from the Dillman Family Association, Family Group Sheet Database, compiled 2026, and from the DFA’s own published articles.