Conflicting records are normal in family-history research. One census implies an 1846 birth. Another implies 1849. A marriage record names the father as Henry. A death certificate names William. An obituary says Scotland while a census says Ireland.

The first job is not to decide which one is right. The first job is to preserve the contradiction.

Do not clean the tree before you understand the problem

Family-tree software prefers one date, one place, one spelling, and one set of parents. Historical evidence often does not.

If four records imply birth years of 1851, 1853, 1854, and 1858, possible explanations include ordinary age reporting error, rounding, a late informant guessing, deliberate age change, a transcription problem, or two same-named people being merged.

Deleting the inconvenient dates removes evidence before the case has been diagnosed.

Build a contradiction matrix

For each issue, record four things:

Issue Evidence to preserve Possible explanations Next test
Birth year 1848 / 1851 / 1853 reporting error; wrong person find earlier-age evidence
Birthplace Ireland / Scotland / Ireland informant error; broader place; identity issue examine earlier records
Father William / David middle name; stepfather; error; wrong person marriage, church, probate
Surname McCarthy / McCarty / McCartney spelling; phonetic change; transcription search observed variants

A contradiction becomes useful when it produces explanations and a test that can separate them.

Treat competing claims symmetrically

Suppose a marriage record names Robert as the father while a death certificate names Thomas.

Do not begin with “Robert must be right.” Write several explanations first: Robert was the father and Thomas is an informant error; Thomas was the father and Robert is wrong; one was a biological father and one a stepfather; first and middle names were used differently; handwriting or transcription was misread; or the records concern different men.

These possibilities do not have to be equally likely. The point is to stop the preferred answer from becoming invisible bias.

Judge information, not just document type

An original record can contain weak information about an earlier event.

A death certificate may be strong evidence that the death occurred because it was created close to that event. Parent names on the same certificate may have been supplied decades later by somebody who never knew the parents.

Ask two separate questions:

  1. How close is the record to the event?
  2. How close was the informant to the information?

Specificity is not the same thing as reliability. A precise town and date copied from an unsourced tree do not automatically outweigh a less precise statement from an earlier, better-informed source.

Check whether both statements can be true

Some apparent conflicts describe the same reality at different levels: Wales versus Glamorgan; a village versus the nearby town used in later records; birthplace versus childhood residence; biological father versus stepfather; maiden surname versus a later married surname.

Before calling something an error, ask whether the two statements could both fit the person’s life.

Watch for conflict clusters

One discrepancy may be ordinary noise. Several discrepancies around the same identity demand attention.

If one Thomas Riley is Irish-born, a labourer, married to Mary, and father of Patrick and Ellen, while a later Thomas Riley is New York-born, a carpenter, married to Margaret, and father of Charles and Anna, the larger question is not “which birthplace is correct?”

It is: Have two people been combined?

That is when the name-variant and identity guide matters.

Four legitimate endings

A contradiction can end in several ways:

  1. One item is demonstrably wrong. The index says 1857, but the image clearly says 1851.
  2. Both statements are compatible. One says Wales, another says Glamorgan.
  3. One explanation is better supported. Several early records support one birth period while one much later source differs.
  4. The conflict remains unresolved. The surviving evidence does not justify choosing.

The fourth outcome is acceptable. An unresolved conflict is better than false certainty.

The wider method in Break Through Your Genealogy Brick Wall keeps contradictions visible through the CHALLENGE stage, then revisits them during CORRELATE and CONCLUDE.