When we find a record, it can be tempting to focus on the one piece of information we were looking for.
- A death certificate that gives us that elusive birth date.
- A census gives us a birthplace.
- A marriage record gives us the name of a spouse.
- Who created this record?
- When and why was it created?
- Who provided the information?
- Would that person have known the answer firsthand?
- Was the information recorded near the time of the event?
- Is the record an original record or a later copy or transcription?
Within any record, information varies in reliability. Information falls into three categories:
Primary information comes from someone with firsthand knowledge of the event. For example, a mother providing information about the birth of her child is reporting something she personally experienced.
Undetermined information is provided by an unknown informant, meaning we cannot assess its reliability. When a record does not identify who provided the details, we cannot determine whether the informant had firsthand knowledge, secondhand knowledge, or no knowledge at all.
Census records are a common example. Most U.S. censuses do not indicate who spoke to the enumerator. For example, it could have been an older child, a second wife, a neighbor. There is no way to determine the reliability of the information given. The only exception is the 1940 U.S. census, where enumerators marked the informant with an X, allowing us to assess whether the information likely came from someone knowledgeable.
A single record may contain both primary and secondary information. A common example would be a death certificate. An attending physician may have had firsthand knowledge of the death or its circumstances. The information concerning the death is therefore likely to be primary information.
The deceased person's birth date, birthplace, or parents' names may have been supplied by a grieving spouse, child, neighbor, or other informant decades after the person's birth. For those events, the informant was probably relying on secondary information.
Then Comes Correlation
While analysis looks at individual pieces of evidence, correlation looks across them and brings the evidence together. It asks:
- Where does the evidence agree?
- Where does it disagree?
- What patterns emerge when records are compared?
- Perhaps the 1882 date appears on a death certificate created 70 years after the birth.
- The 1884 estimate comes from a census.
- The 1885 date appears in a Bible entry that seems to have been written years later.
- But the 1887 church baptismal record was recorded shortly after the child's birth.
Create a Working Hypothesis
Genealogical conclusions should be based on the best available evidence—but remain open to revision if a new record appears. As evidence accumulates, we form a working hypothesis and test it by asking:
- Does the age fit?
- Does the location fit?
- Do the family relationships fit?
- Does the chronology work?
- Do the names make sense?
- Do the records contradict one another?
- What evidence would we expect to find if our hypothesis were correct?
- What evidence would prove us wrong?




