There are several kinds of records that give us a person's age in a
particular year, the most common and obvious being the Federal censuses,
so we are often faced with determining a birth-year from this information.
If I can make no other point on this page than this one, it is that:
| Any record providing a person's
age in a particular year will always produce TWO possible birth
years, depending on whether the person has or has not had their
birthday yet that year. The widespread practice of simply deducting
the age from the year to determine a single birth year is incorrect. |
For example, if someone is age ten in the 1850 Census, it means they
could have been born in 1840 or in 1839, depending on whether or
not they had had their birthday by the time they were enumerated by the
census-taker. The formula is not complicated. Just subtract
the age from the year, as you have always done, but then include the next
earlier
year as part of an age range. |
| If you know the date (year, month, and day, not just the year)
when the person's age was taken, but the age is given only in years,
the calculation is still the same. That is, knowing the precise date
the age was taken (e.g., as in knowing the Census enumeration date)
doesn't change the calculation, at least, not unless you have additional
information (e.g., you know from another source what day the person
celebrated their birthday, but you don't know the year they were
born). |
| If you know the date (year, month, and day, not just the year)
when the person's age was taken,
and you know the age in years,
months, and days (as one often does from a tombstone incription),
then the best way to calculate the birth date is to use Ben
Buckner's JavaScript Birthday Calculator. This derived date may
not be exact, however, because you don't really know if the person who
originally calculated the age did so correctly or used the same method
(there's more than one way to do it). Therefore,
any
calculated date of birth is best preceded by ca. (the
standard abbreviation for Latin circa, meaning "around"). |
| One reason it's important not to imply that you know more than you
do — as in using just a single birth year from census data or leaving the
ca.
off a calculated birth date — is that by doing so you may mislead someone,
including yourself!
For example, if you say your James Brown was born in 1832, instead of
1831/2, someone looking for a James Brown born in 1831 may pass you by
when, in fact, it was a match — just as you, yourself, may be mislead into
not looking for a James Brown born in 1831. Likewise, if you give
a calculated birthdate that is off by a few days or weeks, but don't
precede it with ca., someone else may have the correct birthdate
(e.g., from a family bible), so dismisses your record by failing
to see the match when there really was one. And they failed to see
the match because a full date without ca. (e.g., "12 Oct
1731") implies that you had a sound basis for giving it. In other
words, "12 Oct 1731" doesn't look like a guess or an estimate, it looks
like a certainty. Adding ca. at least lets other researchers
know that there is some "ify-ness" about the exact date.
The bottom line is 1) to not read more into your data than is there
and 2) to always remember a fundamental maxim for any researcher, genealogical
or otherwise:
| Bad data are not
better than no data! |
Don't let your desire to "fill in the blanks" lead you to accepting
more than what is actually supported by the evidence because:
|
