Working the Numbers

The issue with most trees is that they are often uni-directional, we find a set of parents for a known child, then work back to their marriage and work back to those parents births etc. This is filled in by the census to give us some siblings although often not the full picture. However rarely do many people work those trees forwards to find the marriages of the siblings and thus their families on the census, its often more by accident that we find them on the census on the same page but different household to the parents we were looking for or something similar.

The result is that most people have limited numbers of cousins they could enter onto the 1881 census. However as Peter regularly says its these sidelines that hold the key, lets imagine an average 1800s family.

Couple born in late 1700s marry in 1800 and have 6-8 children over course of 20 years to 1820. Of course some of these say 2-3 may die young although experience shows often the numbers are "topped up" so to speak by having families of 10-12 of whom 3-4 die. So I'm going to use 7 children as a first benchmark, naturally some families will be smaller some will be larger lets take that as an average.

So 7 children survive and marry in the 1820s and 1830s, these will have children in the 1820s-1850s and again lets say they have 7 each. This branch is now over 50 people (including spouses), in the 1840s through to 1870s these children will marry and have children lets say 7 each again giving us 343 although of course we will encounter the issue that some for these people who married in the 1840s will have children themselves by the time their cousins are born. So we are probably around 350 people. If these half of these are having 7 children by the 1881 census we are talking 1225 people. However lets trim that right back because some won't be old enough to have 7 children by 1881 so lets just take 2.5 times the children just to trim back the numbers a bit although obviously it could be much more that gives 875 children born between 1860s and 1880s. Throw in the parents of the previous generation still alive and you easily get to 1000 people alive on the 1881 census from just one couple.

Now that is from ONE of your branches who marry in 1800. Typically depending on your age lets assume 65, your parents marry in say 1940s, your 4 grandparents in 1910s, your 8 1g grandparents marry in 1880s, your 16 2g grandparents marry in 1850s, your 32 3g grandparents marry in 1820s. (I've used 30 years each generation there which should usually be more like 25 so I'm going to round up by using 20 years to make first couple in 1800s). Thus your 64 4g grandparents marry in 1800s. So we have approx 32 couples marrying in 1800s.

We saw before that for 7 children average we get 1000 people per original couple in 1881 census, for 32 couples that's 32000 people that could be on the 1881 census from your direct lines. Of course some of these will emigrate some will have less children some will have more - some branches may originate from countries outside the census countries Lost Cousins covers. Some will intermarry thus reducing the numbers as branches converge etc. So the numbers might be less.

Note even if you think 7 children is too high lets feed in 2,3,4,5,6 as possibles this gives:

• 2 children approx 864 relatives to enter to 1881 census
• 3 children approx 2544 relatives to enter to 1881 census
• 4 children approx 5696 relatives to enter to 1881 census
• 5 children approx 10800 relatives to enter to 1881 census
• 6 children approx 18336 relatives to enter to 1881 census

Of course by the time you bring that forward to today's numbers there are hundreds of thousands of people out there who are typically your Lost Cousins. So have a look at your My Ancestors page. Have you really entered all your Ancestors?

 

I won't argue the logic but the numbers seem high and unfortunately reality doesn't match the theory.

But, Peter must have lots of data in Lost Cousins. Assuming everyone adds the Ahnentafel number to LC then some interesting stats could be worked out.

Also, looking at my data in FTA, I have 16 4th GGP's, maybe I nice report could be written to indicate which GGP's are missing parents?

 

When I'm doing calculations of this sort I assume 2 children per generation the adjust the result by the increase in population over the relevant period. However this might not be appropriate for Scotland and Ireland where emigration was higher.

 

They may have gone to the US or Canada in which case they can still be entered, although the chance of a match is lower, especially for the US.

 

The big drag on the numbers that I hinted at but didn't factor in was the incidence of cousins marrying cousins, also the which then dramatically cuts the numbers as you no longer have as many unique individuals to deal with. Also I suspect that which 7 children families on average are common. I think in hindsight I shouldn't have assumed that all 7 then has issue. I suspect using 3-4 having issue is probably closer to being right. In which case this would be equivalent to midway between the numbers for 3-4 children per family.

Anyway the general gist still holds. The numbers get shocking large even for just one branch really quickly. The base point is that if you are just entering direct relatives your matches will be rare on Lost Cousins, you really need to enter all the siblings and ideally their children onto Lost Cousins as it is far more likely that a match will come from a 3rd or 4th generation from a sibling of a 2nd or 3rd great grandparent than it will come from someone who is from your more immediate direct line.

Tim mentions a suggestion for a new report to show the people in your tree for whom you haven't yet found the parents. This already exists in FTAnalyzer and it available on the Treetops tab. The people that appear there are the ones at the top of your tree for whom you have yet to find parents. If you limit this report to direct ancestors then you get a list of the furthest back point on each branch of your tree. It's amazing how frequently we might look at such a list and not recognise the surnames but everyone in that list will be a direct ancestor and finding that birth might just unlock another generation.

I did that looking for a birth for a branch I'd not looked at recently and found a birth fairly easily. I'm not sure why it had eluded me before perhaps because websites have moved on and new info is available and I'd just not looked at that branch for ages. That discovery led to a marriage for the parents which hit a fairly unusual name Moses Arbuthnott. When finding an unusual name my first reaction is to Google it and see if I get any hits. WELL!!! It turns out there is a lot of info on him and his descendants, and indeed the whole Arbuthnott family. There are unproved links between Moses and other branches of the family which if proved would mean that he's a gateway ancestor into the landed gentry and records going back centuries.

Note from one genealogy Moses and his wife have a six children which leads to well over 100 people, and that only includes a very, very small part of my own branch, as you'd expect it concentrates on the person who created the tree's main line.

I've not added in all these people into my tree if I did it would allow me to add probably 50-60 more people to the 1881 Lost Cousins Scottish Census. Just from one couple. Hence relevant to this discussion at hand.

PS. I've moved this thread into its own discussion as it's not specific to the weekly stats Red lists.