* when a source is inconsistent with itself

[jmurphy]

What do you do when a source has information which contradicts itself?

One of my husband's ancestors died in January 1929, allegedly at the age of 84.

A family tradition, repeated in the obit, is that he enlisted in the Army at the age of 18 -- the newspaper says it was on 6 September 1861.

These two sets of ages and dates don't add up.

If I put in the calculated birth year of 1845, that agrees with all the census data I have collected so far. But if I add in the story that he was 18 in September 1861, I am faced with a forest of (!) warnings from Family Historian.

It's certainly possible that he lied about his age when he enlisted, but I have other inconsistencies too, e.g. the year he entered the US. Of course there is no telling who gave those dates to the census taker.

The calculated year of birth, depending on which record you look at, ranges from 1841-1845. And if someone said 'he was x years old when he came to the US so that means it must have been YYYY' then those years would vary also.

I'm cross-checking all the records I can find, but he was born in Canada before civil registration began there, so that won't be any help. (I don't have info on his parents, so haven't investigated church records yet.)

Any thoughts on the best way to deal with this kind of problem?

Jan

ID:3479

[Jane]

I wonder a 6 looks like an 8. So if the only source which is 'wrong' is the army enlistment then I would say that he was probably 16 when he enlisted. My grandfather joined up at 15 and his army book knocks 3 years off his DOB. I have it in as such, so ! on that event, but I just added a note to the event saying that he lied about his age.

[ChrisBowyer]

It's possible of course to record several birth events with different dates, each with their own source reference and assesment.

I prefer in such cases to record one event with what I think is the most likely date taking into account all the available information, probably with an (approx) qualifier, and put something in 'text from source' for each piece of contradictory information, and possibly something in the note to explain why you think it's wrong.

Age at death is in particular often wrong, being of course what his survivors believed. I usually give more credibility to birth dates as supplied by someone's parents where possible.

[jmurphy]

Jane -- it does seem likely that he lied about his age when he joined the service. One muster roll lists him as joining at 18, then a year later, when he was transferred to the regular army, he is listed as 21! (The evidence points to there being only one man, not two individuals with the same name.)

Good points, Chris, of course. The maddening thing about this particular individual is that there are two dates in a single printed (not handwritten) source that are out of range with each other, and the evidence that I have to cross-check which (if either) might be correct also disagrees.

The age at the time of the census is also unreliable, since it's also likely to have been reported by someone else.

I had read case studies about situations like this, where the age is all over the map, so I wasn't unaware that you could see something like this -- it's simply the first time I've seen a case where the variation is so wide.

I would love to get at birth records but so far I don't have the clues to get me there.

Thanks everyone for your comments.

Jan

[jmurphy]

Well, here's an amusing [oops] story.

I forgot that I had a 1900 US Census record for this person.

The 1900 Census lists the month and year of birth, and it matches the month and year listed in the obit.

Once I plug in that birth date, it resolves most of the discrepencies, except for the 'age at enlistment' which as Jane points out, is not likely to be accurate.

In the process of checking this, I've discovered that some of my data lines have the wrong sources attached to them.

Argh.

Jan