I saw this on another list and asked permission to pass it around. This posting
contains the first message sent and a second message showing a faster way of ca.
the ages.
Virginia
Lots of you old-timers are well aware that you can often combine age
data from several pre-1850 censuses to narrow the age range for a
long-ago ancestor who died before the 1850 census came along and got his
actual age. Some new people, however, might not be aware of this
technique. But most of us have a bit of a struggle getting all the age
brackets straight in our heads so that we make accurate refinements in
our ancestors age on the first try. With this in mind, I offer a
relatively easy and foolproof way to make these age refinements.
The first thing you ought to have is a pre-1850 census age bracket cheat
sheet. I use a short document that lists for each pre-1850 census the
census day, the age bracket for each column and the range of birth years
covered by each column in each census. This does away with simple math
errors because its pre-calculated. Unfortunately, the cheat sheet
would lose its formatting if I tried to include it in the body of this
message but if any of you want a copy I will be glad to send you a Rich
Text Format copy as an attachment to email. Or you can easily make up
your own.
Next you need some way to represent in a clear and simple fashion the
possible years of birth for your ancestor from each census. The
clearest and easiest way I know of to do this is to graph the years on a
sheet of graph paper. The graph isnt complicated. Its just a
straight horizontal line drawn from the earliest to the latest possible
year. I cant show graph paper in this message so Ill just show each
possible birth year with an X. If your male ancestor was listed in the
5th column in the 1820 census, then, according to the cheat sheet, he
was born between 1775-1794. This is a pretty big spread. Graphed, it
would be a line covering the 20 years from 1775 to 1794 like this:
1820 census 1775 XXXXXXXXXXXXXXXXXXXX 1794
Now suppose you found him again in the 1830 census and his age was
listed in the 6th column making him born 1790-1800. It is immediately
obvious that he must have been born 1790-1794. This drops the possible
age spread from 20 years to 4 years -- a big improvement. If you
graphed these two age spreads, it would look like this:
1820 census 1775 XXXXXXXXXXXXXXXXXXXX 1794
1830 census ~~~~~~~~~~~~~~ 1790 XXXXXXXXXX 1800
The overlap of the two lines is the new, refined age spread.
This looks almost too simple to bother graphing so lets look at a
little more complicated case.
Suppose your ancestor was 40-50 in 1830, was listed as under 16 in 1790,
and was 16-26 in the 1800 census. What age bracket do you get when you
combine these data? To find out, go to your cheat sheet and find the
spread of birth years represented by each column in each census and
graph the data.
1790 census 1774 XXXXXXXXXXXXXXXX 1790
1800 census 1774 XXXXXXXXXX 1784
1830 census ~~~~~ 1780 XXXXXXXXXX 1790
You can see easily that he was actually born between 1780-1784 and that
he fits the data in the 1790 census as well.
Well, these have been simple examples but more complex examples abound
out there in genealogy. Sometimes you can actually narrow it down to
the exact birth year if everything works out just right. This simple
technique, however, works just as easily on those harder examples.
I hope this helps.
Jerry Merritt
>
>>Yesterday I discussed a graphic method of refining pre-1850
census age
>>data. Some of you asked if there might be an even faster way.
Actually,
>>there is an easier and quicker way. Its not quite as foolproof as
the
>>graph method, however, and its not quite so visually intuitive but it
s
>>really fast. Heres how you do it.
>>
>>You still need to figure out the spread of birth years for each census
>>you have data from -- either from a cheat sheet or long hand -- just
as
>>for the graphic method.
>>
>>Then, instead of graphing the spans of birth years from each census,
you
>>write them down in a column, like so:
>>
>>1774-1790 (from 1790 census)
>>1774-1784 (from 1800 census)
>>1780-1790 (from 1830 census)
>>
>>Now just circle the largest number in the left hand column and the
>>smallest number in the right hand column. In this case 1780 in the
left
>>column and 1784 in the right column -- and thats your refined birth
>>year spread. The person here was born between 1780-1784. You cant
get
>>much simpler.
>>
>>One problem I know of with this method is that you may have an
ancestor
>>whose birthday falls right on the cusp and will be in the next higher
>>bracket in one census and the lower bracket in the next because his
>>birthday came just after the census one time and just before the next
>>time. But you should see that the final numbers are reversed and take
a
>>harder look at the data. If you had graphed a case like this, you
would
>>see that two of the lines on the graph didnt overlap and would know
>>immediately what and where the problem was. Ive only had this happen
>>to me one time, however.
>>
>>The advantage of this method is you dont need to graph the birth
years
>>so its quicker. The disadvantage is that, if one or more of the
>>censuses got the age bracket wrong, it wont be immediately obvious
and
>>could give you an incorrect answer if you werent paying close
>>attention.
>>
>>Whichever way you choose to do it, however, a pre-calculated cheat
>>sheet does 90% of the work for you and eliminates 99% of the errors.
>>
>>Jerry Merritt