I like to use the GEDCOM qualifiers to remind myself of the nature of the approximated birth year. I use "calculated" when the date has been generated by a calculation (back-calculated from a year in a census record). I use "estimated" when I have made some guess of my own (and record the reasoning in a research note). The real challenge comes when forms such as the US WWI and WWII Draft Registrations, which have both a date of birth and an age on them, are not consistent with each other! In those cases, I've found that the day and month are often accurate, but the year of birth has been reported wrongly or calculated from the age. One of the people in my database has a WWII reg card that put his birth in 1892, and as far as I've been able to tell, he was baptized in 1890! All of the records with dates agree that he was born in July, and those that do have a day recorded agree on the day, but the year is all over the place. Jan Murphy On Fri, Dec 25, 2015 at 9:32 AM, Paul via < family-historian-users@rootsweb.com> wrote: > Here are some more thoughts about recording birth dates. > > Perhaps most of the time, you will have several "data points" that imply > birth year - for example one (or probably more) census age, possibly age at > marriage, even age at death. > > In the absence of a definite DoB (usually from birth certificate), the > "real" date will likely have an uncertainty of *at least* a year. > > This is because census age reporting is notoriously variable, and even age > at death may well not be accurate. > > (Incidentally, it is not unknown for claimed birth dates on a certificate > to > be in error - deliberately or not.) > > My own approach is to choose a birth year that best fits the available data > and qualify it with "App(roximate)". > > Note that the Family Historian Property Box Event display is very handy > here > in highlighting which event ages are "out of bounds". > > Note also that I try to avoid a birth date range expression like "Between x > And y" because FH cannot then do its date calculations! > > I don't know how widely this is understood, but apparently FH treats "1870" > or "c.1870" as "1 Jan 1870" for calculation purposes. > > Similarly "Feb 1870" would be taken as "1 Feb 1870" and "Q3 1870" as "1 Jul > 1870". > > The important things are to be consistent about your method of recording > birth date *and* to include reported age in your source citations (so > others > can form their own opinion). > > Just a couple of extra comments. > > There is not *usually*, in my opinion, much point in trying to estimate > birth date to better than a year (some life event ages may be > inconsistent). > > However, with a birth registration index you may be able to narrow that > down > to a specific quarter (e.g. Q3 1870). > > In that case I would choose the birth date as "c. Aug 1870". > > Note that I chose *middle* month of the period on the basis that the actual > *registration* date would be at most about one month in error. > > And in all but exceptional cases the actual birth date would not have been > more than a week or two different from that. > > Happy Christmas, All > > > > P.S. I did not mention the other two GEDCOM date qualifiers "Calculated" > and > "Estimated". > > This is not the place to resurrect discussion about the differences from > "Approximate". > > Suffice to say I get by perfectly well without them. except in some very > special cases. > > > > > ------------------------------- > To unsubscribe from the list, please send an email to > FAMILY-HISTORIAN-USERS-request@rootsweb.com with the word 'unsubscribe' > without the quotes in the subject and the body of the message >