* One or more sources for an event?

[Valkrider]

I am beta testing the new Ancestral Sources plugin and Nick and I had an email discussion about multiple sources for an event.

I am of the view that I want all sources to an event retained in my database, Nick says that that is unusual and most FH users only want one source for an event. [Edit by Nick: No I didn't say that, I said the complete opposite! See my post below]

An example for my reasoning for this is:

Today I find an entry on FreeBMD for the birth of an individual and so I record this with the quarter page etc as a source for an approximate birth year for the individual. Some months later I order the Birth Certificate from this information and I now record this as a source for the actual birth date. I want to keep both sets of information by default so that, for example, if I pass my gedcom to someone else they won't get the image of the certificate but they will be able to check my research and find the entry on FreeBMD.

Unfortunately polls aren't enabled on this forum but what are your thoughts? Do you keep multiple sources for an event or just the last one?

Admin Edit: I have added a Poll for you.

tatewise Edit: The poll has been removed at Colin's request - see post Sun Apr 06, 2014 7:27 am

[DavidNewton]

I voted for it depends because I always source where every piece of information arises, even if the source is an implication from another source (e.g. a census entry). Later when the information is firmed up I will remove this initial suggestion as I feel under no obligation to explain all the routes I take to reach a conclusion.

With regard to one source per event I think that can only hold if your information is original and allows no room for transcription errors (or any other sort of error)

[margarita]

I keep various sources for a single event because they often each contain different pieces of information which build up to the complete picture.
Examples:
BMI birth index containing mothers maiden name
BMI death index containing birth date
Christening record containing date of birth

I would keep all of those sources relating to a birth.

Maggie

[AdrianBruce]

I would keep all - for instance, if I'm trying to work out when and where someone was born, 4 censuses might give me 4 similar dates and places. All 4 would go into the final conclusion, which might match none of the censuses....

Keeping all is particularly true with indirect evidence.

As indicated elsewhere, even with explicit direct evidence, the parts of the "fact" (date, place, address, value, notes, etc) can come from different sources.

[tatewise]

I keep them all, especially where several Census Records indicate Birth details, and the Birth is pre-1837, so finding Church Parish details of Baptism/Christening is more difficult, let alone exact details of the Birth.
The same is true for pre-1837 Marriages, initially indicated by Census Records, with the date inferred from the age of the eldest child.
The date of birth on a Death Certificate would also be an additional Source for a Birth.

Similar arguments apply to other Facts such as Occupation and Residence derived from various Sources that are often all slightly different, so there is no definitive single Source.

[NickWalker]

I am of the view that I want all sources to an event retained in my database, Nick says that that is unusual and most FH users only want one source for an event.

You've misunderstood what I was saying and indeed I would also vote in favour of multiple sources per event - please edit your post because that is most certainly not what I said in my email conversation with you.

Of course you should have more than one source for an event if there is more than one source!

My point was that it is less common for users to, for example, create 3 death facts for an individual, each with their own source. In my experience it is more common for users to create a single death fact and have multiple sources linked to that.

[LornaCraig]

I have voted 'It Depends'.
In many circumstances I keep all sources, for reasons others have given (different aspects of an event, such as date and place, may have different sources; conflicting sources can be used to produce a date range).

However in some circumstances when I obtain what I regard as a definitive source I delete others. For example, if I have a GRO index reference as a source and I subsequently obtain a certificate with more precise information, I delete the GRO index source. My reasoning is that the GRO index is just a 'pointer' to the full entry/certificate, not an additional independent source.

If I have found transcriptions of a parish record of baptism/marriage/burial on various sites I keep them all as sources (particularly if the transcriptions differ) but if I later obtain an image of the parish register page I regard that as definitive. I make my own transcription of the image and include a note if I feel there is genuine doubt about what the entry actually says. I then delete the sources which were transcriptions only. They merely show that someone else had looked at the same page or image, and five people transcribing the same PR entry don't make five different sources.

So it depends whether the sources are genuinely independent of each other. And by deleting 'redundant' sources (those that are not genuinely independent) I reduce the length of the Sources list in reports, which can get unwieldy otherwise.

(Edit: Didn't see Nick's post until after I sent this. No, I would definitely not have multiple birth/death Facts.)

[ColeValleyGirl]

I record all sources for all events with an explanation of how I've reconciled differences in the information provided (if in fact I have been able to reconcile differences).

I may also record finding aids (such as GRO indices) as a transient 'source' but delete them once I've referred to the source they lead me to (such as a certificate).

[AdrianBruce]

Taking what Nick has written - in 99.999% of cases, I would only ever have one birth event (with multiple sources) and one death event (with multiple sources).

The GEDCOM specification (working from possibly fallible memory) suggests that 2 or more birth events should be interpreted as alternatives. I would "never" follow that - I would only ever have one.

However, I am fairly certain that somewhere in my youth or childhood of using FH, I did come across somebody whose evidence was so tangled and fragmentary that it was impossible to reconcile the data into one fact and I just gave up and entered two birth facts. Possibly the reconciled fact in English would have simply said "He was born" (a bit obvious really), whereas the two (alternatives) would read "He was born in Russia" and "He was born in England" respectively - with a note on each saying that they were alternates.

Given that this is so unusual for me, I would be happy to handle that person manually.

Thinking further - I cannot think of anything other than birth or death that would only occur once per person and nothing that would appear only once per family. (Yes, I have two marriage events for one set of my 3G G-parents - thoughI wouldn't trust anything he said!)

[NickWalker]

I think this poll is somewhat pointless (as the results prove). It is obviously the case that multiple sources can be used to support a single fact such as a birth, I was one of the people on this site almost 10 years ago promoting the use of sources (I wrote the original knowledge base article on this) and I developed Gedcom Census to make all this easier. As you could probably tell I was somewhat irritated to have it suggested I thought multiple sources shouldn't be used after all these years of championing sources!

I know some users of Ancestral Sources like to create multiple birth facts, multiple death facts, burials and baptisms where there are differences in what the sources say. Valkrider suggested in an email to me that AS should create a second fact (e.g. death) by default and I simply pointed out that I thought that this wasn't what the majority of users would want. I'd have been happy to have a poll show how common or uncommon that is. e.g. the poll could have said "Do you record more than one birth or death fact for an individual if sources give different information?" and the result of that might have been more interesting!

[Valkrider]

Nick

Sorry that I have irritated you that was not the intention.

We were talking at crossed purposes and I misunderstood what you were saying.

The misunderstanding arose because the current live version of AS does replace some of my existing sources when a new source is added. Maybe I am using it wrongly, maybe I have it configured incorrectly but it is something that I found a problem with and thought with the wording in the beta that it was going to continue doing this. If it isn't supposed to do this I need to spend some time trying to sort out why it is doing this as it is a pain to re-enter a source after AS has overwritten it.

In light of Nick's comments maybe Mike or Jane should close the poll.