Learnings from the Ruby study #1 – Impact of the new GRO index
One of the first things to do when starting out on a new One-Name Study is to construct some core data sets. Apart from being a requirement set by the Guild, there are several other reasons why it makes sense to do this.
1. These lists act as helpful checklists as one reconstructs families
2. They can also be a useful reminder of the scale of the study in different countries and thus possibly aid in decision-making about where to start
3. As one notes which individuals from each data set have been included the notes can be used as a means of checking progress and ultimately for answering the question, “How will you know you have finished?”
The initial Ruby team constructed core data sets for several countries: notably Canada, England and Wales, France, Ireland, New Zealand, Scotland, and the USA. The original England and Wales data set chosen was the 1881 census. We thought it might be interesting to add in births, deaths and marriages from the national indexes. While not absolutely perfect, they are a terrific resource stretching back to 1837 when modern recordkeeping began. This data is, I think, without parallel anywhere in the world.
Learning #1: Technology has had a massive impact in how we work
Paul Featherstone recently related on a Guild webinar how it took him ten years to construct these core data sets for his study from his occasional visits to the repository in London. He started twenty years ago, before the internet and only managed to collect data for years from 1837 to 1900. On the same webinar, I told how doing the same thing ten years ago for 1837 to 1940 took me a couple of days using copy and paste from the FreeBMD website. I then related that “for a new study” (this Ruby Study) it took all of ten minutes to do the same thing for 1837 to 1980.
There are some well-known shortcomings in the original government indexes from which the FreeBMD website was transcribed. The two biggest ones are:
- the mother’s maiden name is not shown on birth records before Q3 1911
- the age at death is not shown for deaths between 1837 and 1865.
Both of these shortcomings have been largely fixed by new indexes issued by the General Register Office. The new indexes also show the full name for every person, though they aren’t as customer friendly to use as FreeBMD.
So, I decided to construct a hybrid index of the birth and death indexes for the Ruby study but adding in the extra details from the new index. (Bear with me a little below because I hadn’t planned to write a blog post and so I didn’t immediately start counting what I found.)
Learning #2: there are more differences between FreeBMD and GRO indexes than I expected
Thus for anyone conducting a One-Name Study with people in England and Wales, you cannot ignore the new index.
To put some numbers on it, now I have blended the indexes there are 624 Ruby births in England and Wales between 1837 and 1911. In the course of blending:
- I deleted 15 duplicate entries in FreeBMD
- there are 11 entries which are in FreeBMD but not in the GRO index
- I added no fewer than 53 new entries. I was honestly shocked at how many there were.
See the two examples in the boxes. Box 1 shows all Ruby births in 1872 from FreeBMD, one boy and three girls. Box 2 shows five female births for 1872 as listed in the new GRO index.
There was a similar pattern with the death entries, with 133 deaths from 1837 to 1865 in the blended index. This number includes 21 entries not in FreeBMD and only 5 in FreeBMD but not in the GRO index.
Now, until we finish building families for England we will not know for sure what the cause of this is. I did observe that there were some registration districts which seemed to be the primary origin of most of the extra entries. These included in alphabetical order: Bolton, Goole, Lancaster, Lincoln, Manchester and, most of all, Wigan.
They are all in the North of England. As a surname, Steve Archer’s Surname Atlas shows that Ruby was relatively heavily represented in the South West and South East of England in the 19th century. Consequently, the increase in numbers from the North of England is even more dramatic than might at first appear!
Maybe there are Rubys in the index who aren’t Rubys, or Rubys excluded from the index altogether.
Learning #3, potentially
So, until we figure out whether the cause is particular to the Ruby name, if your surname is heavily concentrated above a line from the Wash to the Mersey, I suggest you pay particular attention to the new index.
My working assumption is that there are some gaps in the FreeBMD transcriptions for some reason. It could also be that there are several people named Roby (a name concentrated in Lancashire) who have been transcribed as Ruby in the GRO index. One of the extra girls in 1872 mentioned above was a Roby in FreeBMD, for example. Time and analysis will show.
One major caveat to the working assumption is that during the course of adding mothers’ maiden names to the indexes before 1911, I found considerable variation in those transcribed surnames. In fact there were so many variants in mothers’ names for what were obviously the same families it made me doubt the accuracy of the transcription of the Ruby family name itself! So we will need to revisit this whole analysis once we have finished family reconstruction for that time period and have identified every individual named Ruby and those who aren’t Rubys!
from the Ruby One-Name Study Team