Twins puzzle

[Pauline]

Taking the three baptisms example, how do the odds change if looking through the burial register we find that one - or even two - of the candidates died as infants?

I’m not sure that I would see the odds as particularly relevant here. Firstly, I wouldn’t consider any baptism as possible without first checking for a burial and other information.

Secondly, as far as I’m concerned the absence of a burial on it’s own doesn’t make a baptism any more or less likely to be that of my ancestor. It may make the baptism merit further investigation, but no more than that.

 

[peter]

I’m not sure that I would see the odds as particularly relevant here. Firstly, I wouldn’t consider any baptism as possible without first checking for a burial and other information.

That's a luxury that isn't always available to us. Not all parish registers are online, so often we're forced to use indexed transcriptions for our primary trawl, and you'll know as well as I do that baptisms are more likely to have been indexed than burials (for example, there are hardly any British burials in the IGI).

Secondly, as far as I’m concerned the absence of a burial on it’s own doesn’t make a baptism any more or less likely to be that of my ancestor. It may make the baptism merit further investigation, but no more than that.

It's the same thing, surely? We don't have time to investigate everything in detail so we're continually making judgements about what's worth taking a closer look at, and what isn't. Whether we recognise it or not, probability is - or should be - the basis of much of our decision-making.

 

[peter]

Another example is deciding which BMD certificates to order. Even if we're certain that we've found the right index entries, and often we're not, those of us with limited funds have to decide which are most likely to provide us with useful new information.

 

[Pauline]

It's the same thing, surely?

Yes and no. What I’m saying is, that if there were 3 possible baptisms and 2 can be eliminated, then as far as I’m concerned, and regardless of what the statistics might indicate, for the other baptism nothing changes. It remains what it was before - possible.

 

[peter]

Yes and no. What I’m saying is, that if there were 3 possible baptisms and 2 can be eliminated, then as far as I’m concerned, and regardless of what the statistics might indicate, for the other baptism nothing changes. It remains what it was before - possible.

Hopefully once my series of articles is complete you'll look at things differently.

 

[Pauline]

Hopefully once my series of articles is complete you'll look at things differently.

I've no doubt I'll enjoy reading these articles, but I do doubt that you will persuade me to view this remaining baptism as anything other than 'possible' despite eliminating the other two.

I suspect this baptism question is not a good example, though, as very clearly we have insufficient data here to draw any useful statistical conclusions about this remaining baptism. This is often the case with family history puzzles, though I don't deny that we will often make probability assessments without realising we are.

Maybe another thing worth saying here is that we are not all natural statisticians. Despite having A level Maths, having studied maths as part of my physics degree and coming from a family of natural statisticians, I am not one myself. I'm OK with the straightforward stuff and I can understand more complex problems when presented with the reasoning, but calculus was always my thing.

To me family history is a hobby and something I do purely for pleasure, and stressing too much about statistics risks taking the pleasure out of it. What is helpful to some may prove a hindrance to others.

 

[Bob Spiers]

The last time I had my own version of the 'twin lamb' puzzle was in trying to assign a Master to my 18th Century ancestor an apprentice Carpenter. My ancestor finally became a Master Wheelwright and a Guild Freeman with his own Apprentices. But with whom did he serve his 'Carpentry' apprenticeship? The puzzle centred around my ancestor's actual birth year, and explaining how he married within (seemingly) his Apprentice term of 7 years - something quite forbidden by Indenture rules at that time. Finally assigning a Master from 'within my purchased copy of an antiquarian book: "Warwickshire Apprentices and their Masters 1719-1760"

In the end my solution -based on fact and part conjecture - fitted the bigger picture. He served his Apprentice term with an Uncle - a Master Carpenter - and outside of the normal formal Indenture fee afforded only by wealthy parents (even though his father was a Yeoman Farmer), but an Indenture classified as 'en famille'. I also believe he actually started his Indenture at the age of 11 (normally 13) and the year he moved from home to reside within his Master's house. This, tied to his birth year (1720) explained him being out of term c1738; qualified when his Journeyman term began and being free to marry.

It would I am sure have made a puzzle every bit as convoluted as the 'twin lamb' one. I would not normally have gone to such lengths but I needed it to explain anomalies and complete the story I eventually had published in the ' Midland Ancestor' a few years back.

 

[peter]

To me family history is a hobby and something I do purely for pleasure, and stressing too much about statistics risks taking the pleasure out of it. What is helpful to some may prove a hindrance to others.

You only have to look at online trees to see that for many people it's just a bit of fun, and they're not that bothered about evidence.

But for those of us who regard evidence as crucial - and I know you're one of them - it's surely important to focus our limited time and resources on the most hopeful opportunities? I'm not suggesting that we should calculate numeric probabilities - though I know that's something that was put forward in the past when we were discussing online trees - but I think it's important to have a sense of whether and how new information that comes to light should influence how we view the information we already have.

There are plenty of people who only look for supporting evidence, and disregard anything that doesn't fit their preconceptions. Clearly you're not going to find many people like that at LostCousins, but I suspect there are more than a few who might misinterpret evidence - as the confusion over the twins puzzle demonstrates.

So far as the three baptisms example is concerned, that was inspired by a 'brick wall' in my tree - there are three baptisms in the same parish within the space of two years, any one of which could be my ancestor. If I could eliminate two of those baptisms I'd be over the moon.

 

[Pauline]

I suspect there are more than a few who might misinterpret evidence - as the confusion over the twins puzzle demonstrates.

I'm afraid I still don't see the connection. I got confused by the twins puzzle initially but I don't see why this might make me more likely to misinterpret evidence.

So far as the three baptisms example is concerned, that was inspired by a 'brick wall' in my tree - there are three baptisms in the same parish within the space of two years, any one of which could be my ancestor. If I could eliminate two of those baptisms I'd be over the moon.

The point I am trying to make here is that eliminating two of the baptisms doesn't mean the remaining baptism is "the one". That may well seem obvious to you and to other regular forum users, but may not be obvious to everyone reading this.

 

[peter]

I'm afraid I still don't see the connection. I got confused by the twins puzzle initially but I don't see why this might make me more likely to misinterpret evidence.

The twins puzzle is all about interpreting the evidence, ie how should you adjust your expectations when you get the result of the DNA test.

Take as another example the 'brick wall' article in the last newsletter. Although I and several of my cousins matched with the person in New Zealand with Burns ancestors, it turned out that my brother wasn't a match, even though he and I have the same ancestors. Does this matter?

The point I am trying to make here is that eliminating two of the baptisms doesn't mean the remaining baptism is "the one". That may well seem obvious to you and to other regular forum users, but may not be obvious to everyone reading this.

You were considering a different scenario from the one I envisaged - my example was based on my own tree. Perhaps I should have made it clearer that there were only three possible baptisms.

 

[PhilGee]

You were considering a different scenario from the one I envisaged - my example was based on my own tree. Perhaps I should have made it clearer that there were only three possible baptisms.

... and no possibility that your "person" was not baptised?

 

[Pauline]

... and no possibility that your "person" was not baptised?

This was the kind of thing I was thinking of when I mentioned having ‘insufficient data’ in #66.

 

[peter]

... and no possibility that your "person" was not baptised?

The time to worry that our ancestor wasn't baptised is when we can't find a baptism that fits - when we can there's little point worrying about it. After all the chance that our ancestor wasn't baptised, but another child with the same name was baptised around the same time in the same parish is very small.

This is another example of why an understanding of probability is important to genealogists - someone who spends their time worrying things that only happen once in a blue moon isn't going to get very far in their research. The time to worry is when we don't know precisely where our ancestor was born, and we don't have any firm evidence of when they were born.

 

[PhilGee]

The time to worry that our ancestor wasn't baptised is when we can't find a baptism that fits - when we can there's little point worrying about it. After all the chance that our ancestor wasn't baptised, but another child with the same name was baptised around the same time in the same parish is very small.

This is another example of why an understanding of probability is important to genealogists - someone who spends their time worrying things that only happen once in a blue moon isn't going to get very far in their research. The time to worry is when we don't know precisely where our ancestor was born, and we don't have any firm evidence of when they were born.

Well, it's nice to know that two children named Eliza Doel born in Southwick (on all census returns where they appear the birth year is 1806 and both appear in my tree) with only one baptised in North Bradley in 1806 (the parish church) is unlikely to happen very often! The next nearest is 1831. However, my 3gt-grandmother died in 1846 and she was born in 1802 according to her death certificate.

From DNA matches, the two are almost certainly related but I know not how! Just for completeness, nearly all aunts/uncles of "the other" Eliza do not feature as parents in baptism records and a reasonable number of my Southwick "relatives" appear in the Southwick Baptist Register.

 

[Helen7]

After all the chance that our ancestor wasn't baptised, but another child with the same name was baptised around the same time in the same parish is very small.

Is that always the case? You said yourself that you have 3 possible baptisms of one of your ancestors, presumably of the same name in the same parish around the same time. Isn't it perfectly possible for two babies with the same name to be born in the same parish around the same time, with one of them baptised and the other not?

In searching for the baptism of my 3x great-grandmother, I found one that could possibly fit the bill. It was in the right place, with a father of the right name (according to her marriage certificate). However, further investigation of the father (particularly his occupation and marriages) told me this was not the right one. I have still not been able to find her baptism, though other evidence (including DNA) has led me to deduce who both her parents were likely to have been.

 

[Helen7]

Although I and several of my cousins matched with the person in New Zealand with Burns ancestors, it turned out that my brother wasn't a match, even though he and I have the same ancestors. Does this matter?

I would say no. It just means that your brother has not inherited any of the 'Burns' DNA whereas you have. I have come across several cases where I or my husband have a DNA match with a documented distant cousin but not with their sibling.

 

[peter]

Is that always the case? You said yourself that you have 3 possible baptisms of one of your ancestors, presumably of the same name in the same parish around the same time. Isn't it perfectly possible for two babies with the same name to be born in the same parish around the same time, with one of them baptised and the other not?

Absolutely - unlikely doesn't mean impossible. This discussion is reinforcing the importance of being able to estimate probabilities as an integral part of our research.

It's impossible for purely records-based research to be 100% accurate - that's why using DNA to confirm our past research is so important.

 

[peter]

I have come across several cases where I or my husband have a DNA match with a documented distant cousin but not with their sibling.

Whilst what you are describing is technically equivalent, I'm not sure that everyone would see it that way. Whether we're working with records or DNA we tend to give more weight to our own relationships than to the relationships of others.

 

[Bob Spiers]

Although I have been more an observer than contributor in these postings, now it has morphed into to discussing how probability plays its part in genealogical research, I find myself fundamentally agreeing with two of Peter's responses; namely:

#1

We don't have time to investigate everything in detail so we're continually making judgements about what's worth taking a closer look at, and what isn't. Whether we recognise it or not, probability is - or should be - the basis of much of our decision-making.

and #2

This is another example of why an understanding of probability is important to genealogists - someone who spends their time worrying things that only happen once in a blue moon isn't going to get very far in their research. The time to worry is when we don't know precisely where our ancestor was born, and we don't have any firm evidence of when they were born.

They more or less sum up my own approach to Family History research. Take every step you know to ascertain facts, revisit where necessary and when you have exhausted all means known and open to you (DNA included), it is 'make your mind up time'. Record facts, make notes and move on. Don't dwell on the 'what ifs' and 'maybes' and if it all ends in a 'brick wall' return to it later. Remember it is -as Pauline said already - a hobby and should be pleasurable and rewarding.

 

[Pauline]

The time to worry that our ancestor wasn't baptised is when we can't find a baptism that fits - when we can there's little point worrying about it. After all the chance that our ancestor wasn't baptised, but another child with the same name was baptised around the same time in the same parish is very small.

This is another example of why an understanding of probability is important to genealogists - someone who spends their time worrying things that only happen once in a blue moon isn't going to get very far in their research.

I'm afraid I see various flaws in this argument.

Firstly, from my own experience, I think the chances of a baptism which seems to fit not being the right one may be higher than you might think, and I suspect there may be regional variations as well as between mostly rural and mostly urban communities. I have come across plenty of examples where a baptism appears to fit in terms of when, where and even father's name but turns out not to be correct.

Secondly, our ancestors were individuals not statistics. I'm not saying that probability has no place in family history but we need to be wary of assuming things about our ancestors based on majority trends. Personally I would rather check out even a small possibility that a baptism I've found is not the right one than risk climbing up what might turn out to be the wrong family tree.

Thirdly, blue moons may be uncommon but they are not so rare that we can disregard them. More seriously, though, this seems to me to highlight a fundamental difference in approach - is our focus more on "getting on" in our research or "getting it right". For those whose focus leans more towards "getting on" then I can see that weighing up probabilities might be more important than for those of us whose focus is on "getting it right".

Lastly, and slightly tongue in cheek, for those of us to whom statistics do not come naturally, it may prove quicker to check out that baptism than work out the likelihood of it being the wrong one. (And more enjoyable!)