Twins puzzle

[peter]

I think the chances of a baptism which seems to fit not being the right one may be higher than you might think

What does "seems to fit" mean? My example was much more precise than that. What do you think the true odds are in the example I gave?

I do agree that the larger the parish the more cautious we need to be, especially in cities where people moved frequently, and techniques such as family reconstitution can't be used.

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.

Using probability to guide our research isn't treating our ancestors as statistics - it's about using our time effectively, and improving our decision-making processes. No matter how meticulous we are there is always going to be a possibility that we've found the wrong baptism, or that our ancestor wasn't the child of the mother's husband. It hasn't happened to me yet, but until we are able to verify our research using DNA we can never be sure.

is our focus more on "getting on" in our research or "getting it right"

Since we can never be certain that we're "getting it right" until we get confirmation from DNA, all we can do is our best. We can't wave a magic wand and produce evidence that doesn't exist.

it may prove quicker to check out that baptism than work out the likelihood of it being the wrong one

How would you propose to check it out - you make it sound so easy? There's nobody who is going to tell you when you've got the right answer, it's all down to probabilities.

 

[peter]

There are times when we can't avoid using probability, for example when the mother's name is incorrectly recorded in the baptism register, something that's unfortunately rather common.

 

[Helen7]

There are times when we can't avoid using probability, for example when the mother's name is incorrectly recorded in the baptism register, something that's unfortunately rather common.

Can you give an estimate of how common this is? I have come across this, but only rarely, so I wonder if you could put any probability estimate on it. Or does it vary too much between different parishes and different time periods? Say if you found a baptism that was right in all particulars except for the mother's name, it would be useful to know how much this would affect the probability of it being the right baptism.

I have generally looked for other baptisms of children of the relevant parents and made a judgement from these, rather than looking at probabilities as such.

 

[Pauline]

What does "seems to fit" mean? My example was much more precise than that. What do you think the true odds are in the example I gave?

I was actually referring to the example you gave, and from my experience it’s not as as uncommon as you seemed to be suggesting.

it's about using our time effectively, and improving our decision-making processes.

My time is my own and I’m the best judge of how effectively I’m using it. I’m also quite happy with my decision making which has developed with experience.

It’s very clear from this discussion that we have different approaches, and there’s nothing wrong with that. It’s not a case of one of us being right, or doing things better - we go with what works best for us.

 

[peter]

Can you give an estimate of how common this is? I have come across this, but only rarely, so I wonder if you could put any probability estimate on it. Or does it vary too much between different parishes and different time periods? Say if you found a baptism that was right in all particulars except for the mother's name, it would be useful to know how much this would affect the probability of it being the right baptism.

I have generally looked for other baptisms of children of the relevant parents and made a judgement from these, rather than looking at probabilities as such.

Obviously one can never be completely sure that the mother's name is wrong, since there might have been another couple with the same surname and a husband with the same forename - and clearly one also has to eliminate the possibility that the wife died and the widower remarried.

But when I transcribe all of the entries for a specific surname and sort them into families there's quite often an entry that not only sits on its own, and where no marriage or banns can easily be found. There also has to be an appropriate gap in the baptisms to one of the other couples.

Usually - thank goodness - it isn't my direct ancestor whose baptism entry has the wrong mother, but I currently have one that does involve a direct ancestor.

My best estimate of how often this occurs is that overall it's closer to 1 in 100 than 1 in 1000. But I'm sure you're right that it varies between parishes - as you know the baptism register isn't filled out at the time, so it probably depends on whether the vicar keeps notes or relies on the sexton's notebook. Or possibly his memory, though the more detailed the entries the less likely that is.

I don't think I have come across this issue after 1813 when printed registers were introduced.

 

[Pauline]

I don't think I have come across this issue after 1813 when printed registers were introduced.

I have, though it seems to be more common in earlier registers. Quite often it seems to happen when the mother is given the child’s name by mistake, so you get a run of children whose parents are, say, John and Susan, then the next child is Mary daughter of John and Mary, after which it’s back to John and Susan.

I have very occasionally come across it happening with the father’s name.

 

[Helen7]

Quite often it seems to happen when the mother is given the child’s name by mistake, so you get a run of children whose parents are, say, John and Susan, then the next child is Mary daughter of John and Mary, after which it’s back to John and Susan.
I have very occasionally come across it happening with the father’s name.

This post reminds me of a baptism record that has always puzzled me. I have a direct ancestor named James, parents Thomas and Mary, whose baptism I couldn't find, but I found a baptism to the right parents in the right parish at just the right time (late 1805), but the child's name was Thomas. With no later trace of this Thomas, I wonder if this is in fact James's baptism, but in this case the child's name was erroneously recorded as that of his father? I suppose I'll never know for sure, but it must be a possibility (or even probability?)

 

[Pauline]

This post reminds me of a baptism record that has always puzzled me.

And yours reminds me of one where the surname was apparently been given incorrectly - all the details (names occupation, residence) are exactly right, and match that of an older brother 2 years earlier, but the surname is that from the line above. Luckily, although the father is an ancestor this child is not.

Another example is where two unrelated children were baptised on the same day and in the baptism register had been mixed up, each being shown with the other child's parents. This was in 1862 so a family bible and the GRO birth indexes were able to sort out this particular muddle.

but it must be a possibility (or even probability?)

I would agree it must be at least a possibility. Thomas and James can look quite similar at a quick glance, so maybe the vicar misread his rough notes.

 

[peter]

I have a direct ancestor named James, parents Thomas and Mary, whose baptism I couldn't find, but I found a baptism to the right parents in the right parish at just the right time (late 1805), but the child's name was Thomas. With no later trace of this Thomas, I wonder if this is in fact James's baptism, but in this case the child's name was erroneously recorded as that of his father?

At that time it was common for parents to name a son, usually the eldest or second eldest, after the father. So if Thomas is a copying error for James, where is the real Thomas?

Another possibility is that your ancestor was baptised Thomas, but known as James in the family - this would avoided the confusion between father and son. In some families the son would have been called Thomas James and known by his middle name.

Have you checked the BTs? Some errors were corrected when these were compiled (and no doubt some other errors introduced).

Finally, how do you know that James's parents were Thomas & Mary if you haven't found his baptism? I'm sure you have good evidence, but it wouldn't do any harm to take another look at it.

 

[Pauline]

Using probability to guide our research isn't treating our ancestors as statistics

I didn't reply to this yesterday as I was trying to think of the best way to express what I was getting at here.

Probabilities express how things happen across a group or a population (or whatever) and may be less helpful in more individual circumstances. So this Twins puzzle, for example, may be very useful for a sheep farmer with several hundred sheep and give a pretty good guide as to how many female lambs can be expected that year, but may be less useful to a smallholder with just 2 or 3 ewes.

Another slightly different example is one I remember from many years ago when first class mail was expected to be delivered first thing the following day. A survey showed that in our area 95% of 1st class mail was delivered before 9am the next day - but did that mean everyone received 95% of their mail on time, or that 95% of people received all their mail on time, or somewhere between the two? As it happened, we never received any mail before 9am so the 95% figure was pretty meaningless for us as individuals.

So going back to the likelihood of two children with the same name being born in the same place around the same time with one being recorded in the baptism register and the other not, and if (plucking a figure out of the air) the overall odds of that happening were between 1 and 5%, how useful is it to know that? It doesn't really tell me anything much about what might have happened in any particular parish, at any particular time or with any particular individual.

 

[peter]

Probabilities express how things happen across a group or a population (or whatever) and may be less helpful in more individual circumstances. So this Twins puzzle, for example, may be very useful for a sheep farmer with several hundred sheep and give a pretty good guide as to how many female lambs can be expected that year, but may be less useful to a smallholder with just 2 or 3 ewes.

Actually it's the other way round - someone with a large flock of sheep wouldn't spend out on DNA tests because they know that they're going to get a roughly 50/50 split of males and females. But someone with only a small number of ewes might benefit from having a more accurate estimate - it could be a factor in their planning.

So going back to the likelihood of two children with the same name being born in the same place around the same time with one being recorded in the baptism register and the other not, and if (plucking a figure out of the air) the overall odds of that happening were between 1 and 5%, how useful is it to know that? It doesn't really tell me anything much about what might have happened in any particular parish, at any particular time or with any particular individual.

But it does, because until you have better information it's the best information you have. There are numerous reasons why our ancestors' baptisms can be hard (or impossible) to find, and unless we have some rough idea about which of those reasons are most likely to apply it's hard to know how best to proceed.

Nobody can tell me precisely what my chance of dying from COVID-19 is, but knowing the proportion of like me who die is incredibly useful information. There is no difference between using probability to guide our behaviour during a pandemic from using it to guide our research strategies.

 

[Helen7]

At that time it was common for parents to name a son, usually the eldest or second eldest, after the father. So if Thomas is a copying error for James, where is the real Thomas?

Yes, I realise that, and the parents (assuming they were his parents) were married in Dec 1804, making him their first child (born Oct 1805). They had a second son (Henry) the following year, who died as an infant, but I've found no burial record for Thomas junior. I do wonder if the name James was adopted to avoid confusion with the father.

Have you checked the BTs? Some errors were corrected when these were compiled (and no doubt some other errors introduced).

Thanks for the advice. I've looked up the BTs for this parish (Great Easton, Leics) on FindMyPast, but the year in question (1805) doesn't seem to be covered, as it skips from 1721 to 1811 as far as I can see.

Finally, how do you know that James's parents were Thomas & Mary if you haven't found his baptism? I'm sure you have good evidence, but it wouldn't do any harm to take another look at it.

Being a direct ancestor (James was my 3x great-grandfather) I have looked at this several times and came to the conclusion that his parents were Thomas and Mary Freer who married in Great Easton 2 Dec 1804. Freer being quite a common surname in that area (Mary's maiden name was Freer too), I needed to tread very carefully in reaching this conclusion. There are many threads to my reasoning regarding this family that I've pieced together over the years. In summary, my key findings are:
Thomas and Mary had 2 further children, Mary and John.
James married Sarah in 1827 in Skeffington, Leics - one of the witnesses was Mary Freer (his sister or his mother?)
James and Sarah had a daughter Rhoda (not a common name in those parts!) baptised in Great Easton in 1829, followed by 6 further children
In 1841, James' family are near neighbours of Thomas, but Rhoda is staying with the widowed Thomas and his adult daughter Mary (mother Mary died in 1838). This makes sense if Rhoda is Thomas's granddaughter.
The only other Freer family in Great Easton in 1841 are Thomas's other son John with his wife Jane and their son Henry (possibly named after John's deceased brother?)
I also have several DNA matches going back to Thomas's parents as common ancestors, boosting my confidence that I have the right family line.

 

[peter]

Thanks for the advice. I've looked up the BTs for this parish (Great Easton, Leics) on FindMyPast, but the year in question (1805) doesn't seem to be covered, as it skips from 1721 to 1811 as far as I can see.

Bear in mind that whereas register are usually held at the record office that covers the parish, where BTs are held depends on the diocese. So there's just a chance that this is the reason FMP don't have them. You could also check the record office catalogue.

I also have several DNA matches going back to Thomas's parents as common ancestors, boosting my confidence that I have the right family line.

Bear in mind that Ancestry will have based the Common Ancestors matches on what you have in your tree - they wouldn't have looked for alternative solutions.

 

[Pauline]

Actually it's the other way round

OK, so maybe this wasn't the best example to use, but it still makes the point that probabilities have a differing usefulness according to the size of the sample.

But it does, because until you have better information it's the best information you have.

That's partly my point - it's may be the best information you have, but that doesn't necessarily make it good enough to build a tree on.

In my postal services example, someone may have thought, on the basis of the 95% figure, that it was a reasonably safe bet to post me something that I needed before 9am, whereas actually it had zero chance of arriving on time.

 

[peter]

OK, so maybe this wasn't the best example to use, but it still makes the point that probabilities have a differing usefulness according to the size of the sample.

Probability is a factor in both cases, it's the action that is taken that differs.

That's partly my point - it's may be the best information you have, but that doesn't necessarily make it good enough to build a tree on.

Exactly what I've been saying - whether it's good enough to put in your tree will depend on what the probability is.

This needs to be balanced against your subjective estimates of the risks of:

(a) researching the wrong line, and
(b) not researching at all because the 'brick wall' remains in place

For some people (a) will be their worst nightmare, for others it will be (b). Most of us are somewhere in between.

When I'm faced with these sorts of decisions they are numerous other factors that come into play, eg:

• Are there likely to be new resources becoming available that might make a difference? For example, I deferred making decisions about some my Suffolk ancestors until I was able to purchase the relevant CD ROMs from Suffolk FHS
• How likely is it that DNA will provide a solution? This is mainly about how many generations back the baptism is.
• Does the relevant baptism appear in any online trees?
• How difficult will it be to research further back eg if I accept the baptism will I simply be exchanging one 'brick wall' for another one?

 

[peter]

In my postal services example, someone may have thought, on the basis of the 95% figure, that it was a reasonably safe bet to post me something that I needed before 9am, whereas actually it had zero chance of arriving on time.

The problem would not be with the statistics, but the fact that you failed to communicate with the sender. With no other information available sending the item by post would have been the right decision 95 times out of 100.

 

[Pauline]

Exactly what I've been saying - whether it's good enough to put in your tree will depend on what the probability is.

No, that's the bit I'm disagreeing with - bearing in mind we've been talking about overall probabilities here.

I agree more with what you go on to say after this, but I think we may be going round in circles here so perhaps this would be a good time to end this discussion.

 

[Helen7]

Bear in mind that Ancestry will have based the Common Ancestors matches on what you have in your tree - they wouldn't have looked for alternative solutions.

Yes, I'm aware of that, which is why I check the lineage of the DNA matches against documentary records to see if they fit. It's just another piece of supporting evidence in any conclusions drawn.

 

[Pauline]

The problem would not be with the statistics, but the fact that you failed to communicate with the sender. With no other information available sending the item by post would have been the right decision 95 times out of 100.

OK, so maybe once again I didn't pick the best example as illustration, but I think you are obscuring the main point of what I'm trying to express here by focussing on that instead.

 

[Helen7]

Does the relevant baptism appear in any online trees?

In the case of my ancestor James Freer, he appears in over 30 public trees on Ancestry. None have his baptism, but the majority have his parents as Thomas and Elizabeth, not Mary. This appears to be based on the fact that in 1841, Thomas is living with a woman of similar age called Elizabeth Freer, leading many people to assume that she is his wife. However, Elizabeth is listed below his daughter Mary and moreover has 'Ind' under occupation, which I believe means 'Independent'. Correct me if I'm wrong, but I believe this means she is not his wife but of 'independent means'.

This confused me for a while too, especially as I couldn't find a marriage for Thomas and Elizabeth (nor any children for such a couple). Then I realised that she could well be Thomas's sister-in-law, the widow of his late brother (I had traced his brother Robert's marriage to Elizabeth King and Robert's demise in 1836). Thomas and Elizabeth may well have been an 'item' in 1841 of course, but would not have been legally able to marry at that time due to the ban on a man marrying his brother's widow.