Continuing Research When Records Disagree

Define Exactly What Conflicts

When two records disagree, do not immediately delete one. Write a specific question, such as whether Wang Jing was born in Dong Village in 1912, rather than saying vaguely that grandmother’s records are wrong. Separate names, dates, places, and relationships. Sometimes only the birth year differs while other clues agree; sometimes two namesakes have been merged. A clear question makes the relevant evidence easier to see.

All people and events below are fictional examples. Suppose a later-life registration gives Wang Jing’s birth as 1912, while her age in a school register suggests around 1911. Record each original statement and show that the second date is your calculation, not a year explicitly written in the register. Retain “around” and do not average the dates. A one-year difference may involve an approaching birthday, copying error, or another cause; each explanation needs checking.

Compare Sources and Informants

A source is the item you consulted, such as an original register, its scan, or a later family history. Information is a particular statement within it and may come from the person, a relative, or someone unidentified. An official form can contain a birth date supplied from family memory decades later; a private letter can describe a recent event directly. The labels “official” and “family tradition” alone cannot determine reliability.

For each disputed statement, note its creation time, purpose, possible informant, and known copying history. In the fictional example, the school register was made at enrolment and the later registration decades afterward, but a school age might have been adjusted to satisfy admission rules. Proximity to the event matters without automatically deciding the issue. Marking an informant unknown is more accurate than imagining a witness for the record.

Check Whether It Is the Same Person

If identity is unsettled, comparing dates may be meaningless. Examine addresses, spouses, occupations, parents, neighbours, and signatures together for combinations linking records. A common name is insufficient, and one matching detail may be coincidence. A move or occupational change need not indicate another person; seek evidence connecting the before and after rather than expecting an unchanged life.

In the fictional example, both records for Wang Jing name husband Zhao An, but one gives Dong Village and the other South Street. A moving letter explicitly naming both addresses could support the link; merely knowing the villages are nearby does not settle identity. List clues supporting one person and clues suggesting another, and check their independence. Keeping two candidate people can save more rework than forcing every record into one file.

Do Not Count Repeated Information as Separate Proof

Five online pedigrees with the same date may all copy one tree. A newspaper obituary and later memorial booklet may both depend on one relative’s account. Find the earliest material they cite and sketch a source chain: who obtained information from whom, where new information entered, and where copying occurred. The number of sources differs from the number of independent accounts; repetition does not make an uncertain statement reliable.

Write each explanation as a hypothesis that could be disproved. “The school age was reduced by one year” requires admission rules or related records rather than being established by the conflict itself. Consider date conversion, different calendars, and event dates versus registration dates when the material suggests those issues. Do not hide missing evidence behind a long list of possibilities; prioritise a record that can distinguish the explanations.

Design the Next Search

The next step should target the gap rather than expand to every archive. For a disputed birth year, seek earlier school, marriage, or family records; for a disputed place, look for address continuity. State what a new record should answer and what result would change your judgment. Document unsuccessful searches and their boundaries: years covered, name variants used, and regional coverage. A failed search does not establish that an event never happened.

If a search of a school’s 1920–1923 registers does not find Wang Jing, it establishes only that she was not found within that search, not that she never attended school. Record incomplete surviving registers as a limitation. Reasonable research covers material relevant to the question, but there is no fixed threshold of “three sources are enough.” Sufficiency depends on identity links, information quality, conflict resolution, and significant searchable gaps.

State the Conclusion and Its Limits

Finish with a paragraph explaining the most reasonable current answer: supporting material, opposing material, how you compared it, and unresolved parts. A conclusion may be “Born around 1911–1912; the exact year remains uncertain.” This is more useful than selecting a tidy date because the next researcher can see the problem. Evidence should determine confidence, not your desire for a complete story.

Keep a short explanation on the person’s page and make the detailed analysis findable. If clearer evidence later appears, update the conclusion and revision date while preserving the previous judgment and reason for change. Family history does not require solving every puzzle; it requires treating the people behind records seriously. Patience with conflict protects facts and keeps us from shaping a life into a story chosen in advance.

A small comparison table can give each statement a row and columns for source, creation time, informant, identity link, and limitations. Do not assign arbitrary scores and add them mechanically: one clear relationship statement is not equivalent to five copied dates. The table reveals gaps; a conclusion still requires prose explaining how evidence supports or opposes other evidence and why one explanation is currently more reasonable.