Year of Birth etc

makfai · Post by **makfai** » 19 Mar 2009 14:08

I have only just started to try GC (I know - my fault!). I have 600+ individuals so I wanted to test it before using it. I have come across an issue which may indicate that I am doing something wrong and wold welcome advice.

It has been my system to insert a Year (e.g. '1861') only if I know for sure that this is the correct Year. Any other entries will be qualified e.g. '1861 (app)' or 'between' etc.

I have just used one of my ancestors and his family as a test for GC and note that the auto-entry based on age does not put a qualifier in such as 'approx'.

It does not seem to reflect the overlap of years. For example, a person in the 1911 Census (conducted on 2 April) listing an age of 11 can be born anytime between 3 April 1899 and 2 April 1900.

Would it not therefore be more accurate if the correct entry was:
1. 'between' those exact dates (or years and months); or
2.'1900 (app)'.

At the moment the Year which is auto-inputted does not indicate that it is not an exact entry. This could cause confusion and when a report is printed gives an inexact date.

ID:3526

jmurphy · Post by **jmurphy** » 19 Mar 2009 15:54

This is my standard practice -- others may do something completely different.

I generally go in after I use Gedcom Census and add other data (modern US censuses often have more information than the UK censuses record). If you wish to have full year ranges rather than a calculated date, you can always change this once the birth event has been created by GC, or put that information in a note.

I use 'calc' for dates generated by Gedcom Census, for dates that I calculate myself with Excel, and for any dates generated by a calculation by a website for my convenience (Steve Morse's One-Step Web Pages, Ancestry, etc.).

I use 'Approx' for dates of residence taken from sources like City Directories, where the lead time between the collection of the information and the publication of the directory is not known.

I use 'estimated' for things which are my guesses and put my reasoning in a note.

Thus in cases like this, if I have not put in the full year range, I would always use 'calc' rather than 'approx'. Using 'calc' is a visual reminder to myself that the date represents a range 'between [start date] and [end date]' rather than the exact calendar year.

Hope this is helpful.

Jan

makfai · Post by **makfai** » 19 Mar 2009 16:36

I avoided 'calc' for one reason only in this context - if I had calculated it then the result would have been between 2 years not just one year.

The prob with this is when you enter 'between' in FH Date Assistant it shows in the Record Page as 'aft' (i.e. 'after' not 'between'). So if you try to enter 'between 1 April 1884 and 31 March 1885' the info on the Record Page is displayed as 'Aft 1884'. If it was to show up as it does on the Property Page of FH (as 'bet') then that would be fine. In fact I don't know why it shows up as 'aft' when there is a specific option for 'after' in the Date Assistant. [I have raised this as a topic on the FH Forum.]

So in the end I decided to use 'app' (for 'approx') so that (as 'calc' acts as a visual reminder for you) 'app' would alert me to the fact that if I have just entered the one year it should not be taken at face value.

Interesting you referring to an Excel 'Converter' I often wondered if a basic calculator of age into possible DOBs could have been included in FH. An auto-calc for age is included in FH.

jmurphy · Post by **jmurphy** » 19 Mar 2009 19:37

There is no 'Excel converter' as such -- I was referring to the simple worksheets I make for myself in Excel.

If I see in the census 'number of years married' I have one column for the census year, one for the number of years married, then in a third column I put the formula (subtracting the years married from the reporting year).

Obviously this calculated year is subject to the same problems one finds with the year of birth calculated from age. When a couple says in 1900 that they have been married 6 years, it is easy to do the math and arrive at the estimate of 1894.

On the other hand, when the census question is 'age at first marriage' things can be a little more complicated.

With a pre-made worksheet, you can put in notes reminding yourself which census asked 'number of years married' and which asks 'age at FIRST marriage'. Confusing the two questions can lead to hilarious results.

Jan

makfai · Post by **makfai** » 19 Mar 2009 20:03

I understood you had created it yourself. I had thought of doing it myself so that I didn't have to do mental arithmetic each time!

I suppose we are all in the same boat and that is why I wondered why there was no 'calculator' for this in FH. It will auto-calc age from dates but not vice versa.

Of course, a simple subtraction gives one alternative but with the UK Census being in March/April there are always two options for what year a person was born.

Post by **NickWalker** » 19 Mar 2009 22:57

.

have just used one of my ancestors and his family as a test for GC and note that the auto-entry based on age does not put a qualifier in such as 'approx'.

That isn't correct. Gedcom Census does put in a 'calculated' qualifier. This is similar to an 'approx' but shows that it was approximated based on a calculation. A between date can't have a 'calc' 'approx' or 'estimated' after it so putting a between date can make it appear as if you are certain their birth is between those given dates whereas we all know the age on a census is frequently wrong so this could be misleading. Ideally it would be a between date with a calculated on the end but the GEDCOM standard used by FH doesn't allow this. Therefore I think the calc is the best option open and the least misleading. The 1841 census (where age rounded down to nearest 5) will create a between date but again I wish I could have a 'calc' after this.

makfai · Post by **makfai** » 20 Mar 2009 08:58

I am not dissatisfied with the 'calc' qualifier but the test case I ran didn't put a qualifier in. I must have done something wrong - perhaps there was already a year in the sample I was using. I will try another test.

On a broader matter I also wonder why the calculation results in a year which is simply the Census year minus the age. I am as guilty as the next in doing this even not using GC. However, my action is not founded in probability for a UK Census which is usually conducted around the start of the second quarter of a year.

If we take the example of a child whose age is 10 in the 1911 Census then (if the age is correct) the child was born anytime between 3 April 1900 and 2 April 1901. This means that - as we usually default to taking the figure 10 from 1911 resulting in a year of 1901 - there is actually ONLY a possibility of the child being born in the first quarter of 1901 whilst there is a likelihood (which is greater) of the child being born in one of the the last THREE quarters of the year 1900.

By applying this simple arithmetic we end up displaying a year which has 25% likelihood of being correct compared with 75%. So if we were to calculate the likely year based on the probability suggested by the months in which the child could have been born then we would need to start subtracting the age +1 from the Census year.

All seems so simple when you enthusiastically start to bring order to your records!

I suppose as long as each of us (and the people who may later use our data) know what our figure represents then that is OK!

Post by **NickWalker** » 20 Mar 2009 12:29

Yes I know what you mean and this is a fair comment and something I am certainly well aware of in my own research. My reasoning at the time was that this was what Ancestry did so it makes it easier when Ancestry ask you to enter an estimated birth date if it is consistent with what's in FH.

In the next version I might look at making it subtract a year from the dob if the census month is earlier than July.

makfai · Post by **makfai** » 20 Mar 2009 20:27

Please understand that this is NOT a complaint. You have done a splendid job already and the result produced now is probably the one we all would expect to see.

It 'drove me daft' when I first realised the situation created by the Census year vs the age. I was confusing myself trying to calculate the 'between' dates...'should it be 1 April or 2 April?' After a time I simply resorted to taking the age away from the Census year and inserting it in the approx part of the Data Entry Assistant (so nothing new there then!)! The concept being that this would alert me and others to the fact that the date was not accurate. I now just add a Note on what the approx means i.e. age stated at time of Census.

So I did not start this to advocate a change - I just wanted to 'air' the matter. I had wondered whether there were, among genealogists, accepted international conventions for recording data where this ambiguity exists? I am too new to all this to know about such things[confused]

ChrisBowyer · Post by **ChrisBowyer** » 21 Mar 2009 07:28

Whether there's an accepted convention I don't know (but I doubt it).

The problem it seems to me with worrying about the difference between approximate, calculated, estimated and so on, is that more often than not the date of birth is actually a probability, based on a variety of sources such as date of baptism, conflicting census ages, age at death, at marriage, or whatever else you can find.

Unless we have an actual date, we just record everything as an approximate year unless we've found a GRO birth index, in which case it's an unqualified year unless it's first quarter (due to Christmas of course, many December births are registered in the first quarter next year).

For census ages we find it easiest to stick to census year minus age, with a mental note that it could always mean the year before (and often the year before that... I'm sure I've said before on here somewhere, ask any woman how old her children are and you'll often get 'he'll be 10 this year'. In semi-literate early 19th century communities, that's often what gets on the form).

For GRO index dates, when we started we recorded the quarter date, but came to the conclusion that it implies a spurious accuracy... He wasn't born between April and June 1900, that's when it was registered. But it does mean he was probably born in 1900, and after all, nearly everything on the tree is to some extent a matter of probability.

But of course, put everything in the sources (we use text from source and citation notes more or less arbitrarily for this) so you can revise your estimate when another bit of conflicting information arises.

jmurphy · Post by **jmurphy** » 21 Mar 2009 17:28

Chris, to follow up on one part of your comment --

For the GRO index dates, I have recently made custom 'birth registration' 'marriage registration' and 'death registration' events to hold the date of the index entry.

If I have no other date for the event, I may also put the quarter date into the field for the ordinary BMD as a placeholder. Once I have more accurate information, I can delete the placeholder and leave intact the registration event with the index details.

I did this by analogy with the christening dates, for which everyone knows, the event happened sometime after the baby was born.

Jan

makfai · Post by **makfai** » 21 Mar 2009 18:25

I wondered if there was a facility in FH (like an Introduction/Methodology) area where we could say what our own methodology is.

We of course know what it is but those who may be reading our production after we are no longer around to consult may not be so clear.

ChrisBowyer · Post by **ChrisBowyer** » 21 Mar 2009 18:54

There is a Header Record in the Gedcom that you can get at in FH with View > Special Records >. It's originally intended for submission to the LDS (and probably only there in FH to maintain 100% compatibility), but it contains a free text File Description in which you could reasonably record that kind of stuff if you want to.

makfai · Post by **makfai** » 21 Mar 2009 18:58

That may be a useful tip. Thanks!

treefrog · Post by **treefrog** » 23 Mar 2009 10:17

You can always look everyone up in FreeBMD and record the quarter when their birth was registered - this is a bit more accurate (except of course where a family has done a batch baptism, as one of mine did - 6 at once!). Then the date will show as Q1 1888 or whatever, and that'll remind you that it's not a precise birth date.

Jocelyn

makfai · Post by **makfai** » 23 Mar 2009 17:58

Useful suggestion thanks.

Only problem I have with some of the people I am researching is that they are not always locatable in the Births section even when (supposedly!) I have their name and age.

mezentia · Post by **mezentia** » 03 Apr 2009 07:30

When I record a date of birth based on the registration date, I always use Q1, Q2, ... plus the year, e.g. Q1 1901. This method has two benefits. First, it identifies the fact that the birth was registered in that period and might actually have occured in the previous quarter, and second, that there is a birth certificate available which I have yet not obtained, otherwise there would be an actual date.

There will always be anomolies,however. I have one birth certificate recording a birth date a few days after a census, but the child is on the census as 6 months old! The birth was registered in Monmouth, but the family were living in Leominster on the census day, and the enumerator listed the child's birthplace as Aston, Lancs., which was consistent for all the following censuses. To add to the mystery, the child's surname on his birth certificate is different to the surname the family used on the census.

David

ChrisBowyer · Post by **ChrisBowyer** » 03 Apr 2009 09:00

David Dewick said:
the child's surname on his birth certificate is different to the surname the family used on the census.

David

We have a lot of those where the birth is registered with the mother's maiden name and they were married soon after, so the child always used the father's name.

treefrog · Post by **treefrog** » 03 Apr 2009 09:52

Talking of births...Ancestry now have searchable births from 1837-2005. Presumably they're working on the marriages and deaths too.

Glad I renewed my subscription!

Jocelyn

Family Historian User Group

* Year of Birth etc

Year of Birth etc

Year of Birth etc

Year of Birth etc

Year of Birth etc

Year of Birth etc

Year of Birth etc

Year of Birth etc

Year of Birth etc

Year of Birth etc

Year of Birth etc

Year of Birth etc

Year of Birth etc

Year of Birth etc

Year of Birth etc

Year of Birth etc

Year of Birth etc

Year of Birth etc

Year of Birth etc

Year of Birth etc