1939 Register up and running (Part 2)

[Aulus]

Thanks for reposting this here Andy: this is really interesting.

I don't suppose it'll stop people moaning about FindMyPast, sadly.

Hi

The 1939 Register is probably only accessable to a small number of people at Findmypast as anybody working on the Register is required to have a security clearance (from Tender Document).

As Guy has mentioned transcribing was done in columns not rows(as would be the norm) so transcribers did not see a whole line let alone a page.
 
Unlike other releases it is likely that The National Archives dictated what could be released and how so I think blaming Findmypast without knowing the facts is a bit unreasonable. Whilst it seems logical that all addresses should be included even if there are no unredacted occupants the instructions may not allowed for this.

There was a post on a Rootsweb list regarding the Register that is worth reading.

Dear all,
Between 1987 and 1990 I spent a lot of time working with, what was by then
called, the NHS Register.  The discussion about computerising patient
records had begun and I spent some time explaining the need for a unique
identifier and pointing out that the current NHS Numbers would not work. (I
will not give you're the technical details!).

The 1939 Register was used as the basis for the NHS Register.  The records
were kept up to date because they showed which patient was allocated to
which GP.  When a patient died the GP was supposed to return their medical
file and as a consequence would have their capitation fees reduced.  (You
might see a reason there why some of them did not report deaths.) If a
patient died in hospital, or in other circumstances, the GP might not have
known about the death and the patient would remain on their books until
there was some kind of audit (a rare occurrence!).  The level of inflation
in capitation fees was estimated as being between 10% and 30%.  No one would
do anything about it for fear of upsetting the doctors.  (When the NHS
started someone asked how the doctors had been persuaded to join and was
told by the politician in charge "We stopped their mouths with gold!")

Linking death registrations to the register may have been attempted but it
is fraught with problems.  Deaths are often reported by people with little
real knowledge of the deceased.  (My grandfather registered the death of
someone who had taken in grandma when her parents died and reported the age
as 101.  He would have had no idea of her age.  She might have been about
80!)  Doctors and care workers often know only what the person has told them
and exaggerating you age is not uncommon. For married women there will be no
clues to their birth name and lots of people change or misspell their own
names.  You only have to look at the census records to see how mangled a
name can get when the recorder has only heard it and the speaker cannot
read.

My expectations of this new data set were pretty low.  It might help you
find A date of birth but I know all of my direct family ones for this period
from Grandma's Birthday Book.  Bear in mind that this set suffers from the
transcription errors of the original recorders as well as those introduced
by FindMyPast.

I hope this helps.


Andy

[smudwhisk]

Do you mean "security clearance"   or do you simply mean   "confidentiality declaration"

Are people Positively vetted?   I doubt  that they are?

If they have to have security clearance, depending on what level, it can even involve looking at the backgrounds of your immediate family in detail, including financial records.  There are different levels depending on what job you are doing.  Its far more rigorous than just Enhanced Criminal Records Board clearance (or whatever its now called) which I had a few years back working in the NHS even though I had no access to any patient records.  I've several friends who have security clearance, one works at an airport and the others in IT for organisations providing services to government or public sector organisations.

[Aulus]

Guy, I understand your points in defence of FindMyPast, but surely with such a major piece of data and considering the undoubted costs involved, they could have taken a little more time and done some analysis (perhaps using people like yourself and others experienced in genealogy) in order to identify topics and pitfalls that might show up when the public were allowed to view it. It all seems a bit rushed to me - returning to a point I made earlier, why are complete lines redacted and not just the personal details - someone didn't think it through before release

Steve

Of course, if they had taken more time over it, we'd all be here complaining about how long it's taking them, and why can't they release their interim transcriptions so we can see progress. 

Regarding the redaction of whole lines including addresses, I suspect this will again be down to the terms under which FindMyPast are working, imposed by our beloved (sarcasm) government and its mandarins.  I can - unfortunately - also see a real reason why the whole line has to be redacted.  If you look at some of the images of pages from the Register, there are later "notes" added to the line wherever there was space, and on most records, the address column had the most space as for the majority of records the most it contains is "ditto."

What would be really cool is if FindMyPast transcribed the 1939 electoral rolls (even if they just used OCR initially) and cross referenced them to the 1939 register.  The redacted entries on the 1939 register could remain redacted, but the names and addresses from the electoral rolls could be used to search.  The electoral roll is a public document, not subject to the Census Act or the same data protection restrictions as the 1939 register.

I've already used the 1939 electoral roll (on Ancestry) to find the address for someone whom I couldn't find on the 1939 Register, and having the address, I could find them - not that I can understand from the index entry why I couldn't find them by name!

[Aulus]

That being said, if Findmypast were aiming for an accuracy rate of 98.5%, I think they've failed.

If the population of the UK in 1939 was about 46.5 million (a quick google - may be inaccurate), then 1.5% of that would by 697,500.  That's a lot of errors within their contracted margin.  And I'm sure I've read that that was 98.5% of readable records.  If there were a million records with inaccurate transcriptions, that's still just a 2% failure rate.

Transcribing by column must have made an often difficult job even more challenging.

[clairec666]

Of course, if they had taken more time over it, we'd all be here complaining about how long it's taking them, and why can't they release their interim transcriptions so we can see progress. 

True Admittedly I was itching for the register to be released and (silently) urging Findmypast to get a move on!

[clairec666]

If the population of the UK in 1939 was about 46.5 million (a quick google - may be inaccurate), then 1.5% of that would by 697,500.  That's a lot of errors within their contracted margin.  And I'm sure I've read that that was 98.5% of readable records.  If there were a million records with inaccurate transcriptions, that's still just a 2% failure rate.

Fair point... I'm just judging on how many mistakes I've found already. Seems nearer 5% to me? Then again, some of them may be due to bad writing, or mistakes on the original copy, not the fault of the person transcribing.

[Guy Etchells]

Of course, if they had taken more time over it, we'd all be here complaining about how long it's taking them, and why can't they release their interim transcriptions so we can see progress. 

Not only that but it would have cost them far more and that cost would have had to have been passed on to us the users.
Cheers
Guy

[dawnsh]

Currently having fun at TNA doing your research requests.

With regards to my own research today, with one exception, I've reported spelling mistakes for correction on every entry.

Having seen a wide variety of schedules, with amendments in coloured ink, overlapping details and fading ink, I'm not surprised the transcribers had problems. Only one schedule was written in capital letters and that was transcribed correctly.

[ReadyDale]

That being said, if Findmypast were aiming for an accuracy rate of 98.5%, I think they've failed.

If the population of the UK in 1939 was about 46.5 million (a quick google - may be inaccurate), then 1.5% of that would by 697,500.  That's a lot of errors within their contracted margin.  And I'm sure I've read that that was 98.5% of readable records.  If there were a million records with inaccurate transcriptions, that's still just a 2% failure rate.

Not disputing the general theory of what you say, but from that (Googled) figure of 46.5m, you must subtract those in Scotland and N.Ireland not covered in the FindMyPast release.
FindMyPast themselves are quoting 41 million.