1921 census review

[Thunder]

An interesting fact came to light today, after discussion with FMP over a problem with buying an original census record. The 1921 census was transcribed in India. Presumably a matter of cost.

The GDPR (General Data Processing Regulation) doesn't apply outside Europe. The reason why FMP may be cost but in my view is wrong, so no-one in the UK can do it?.

[Guy]

An interesting fact came to light today, after discussion with FMP over a problem with buying an original census record. The 1921 census was transcribed in India. Presumably a matter of cost.

The GDPR (General Data Processing Regulation) doesn't apply outside Europe. The reason why FMP may be cost but in my view is wrong, so no-one in the UK can do it?.

No but it does apply to the National Archives who could not send copies of a closed census to India to get round the GDPR.
Cheers
Guy

[AntonyM]

An interesting fact came to light today, after discussion with FMP over a problem with buying an original census record. The 1921 census was transcribed in India. Presumably a matter of cost.

Which would be expected - transcribing massive amounts of data like this just isn't economic to do in the UK. I was party to similar projects of transcription of government (policing) data some years ago in a previous profession. Apart from the much lower cost, there was always an argument presented that non-native English speakers will transcribe what is actually written, rather than the temptation to write what you think it should say ( there are arguments for and against that).

The normal requirement to mitigate that would be for each piece of data to be transcribed independently by 2 (or even 3) different workers, and where they produce different results it would be looked at again by a more experienced (and probably UK based) reviewer to decide the final transcription.

FMP today seem to have admitted they haven't done the required quality control before the release.

“Due to the secure nature of the 1921 Census project, the period of time in which we have been able to access and review the data ahead of launch has been limited.

“As a result, we have been unable to conduct the same level of quality assurance checks we would normally apply to such a major release meaning you may spot more transcription errors than usual.

“However, to ensure the highest possible transcription standards we will be continually reviewing the data to correct any and all errors over time.”

https://www.whodoyouthinkyouaremagazine ... on-errors/

[Thunder]

TNA is unlikely to have sent the data to India but it would have been FMP. In my view the transcription is not a surprise to me as a number of months ago it was admitted by TNA at the User Advisory Group that the transcription would need to be done again. We should not forget that the award of the contract by TNA was three months beyond where it should have been and TNA have admitted it was a mistake to award FMP the 1939 Register work, so they give the Census work to them!. It is not a question of interpretation of the data but not including the data on the images, e.g. 'born in England', 'born in London' and in a large amount of data cases there are no names, no ages, no places of birth, just that they were in the Army. TNA have, once again, thrown the work to the private sector with no care for researchers, just to make money for TNA. I doubt that any of the transcriptions have been seen by a UK-based researcher, for example most people know that Salford is near Manchester and in Lancashire not in Yorkshire as transcribed!.

[AdrianBruce]

Tongue in cheek: Never mind indexing what you see, what about when the enumerator has amended the householder's text?

For instance, on the very first image that I downloaded, one of the sons was described as working for "Lord Crewe's Estates" - that's been amended to (working for) "Marquis of Crewe's Farm" (at) "Crewe Hall". I can see the original there and personally, I feel certain there are no issues with those changes.

Another son, however, is described by the enumerator as working for "2nd Battalion, Cheshire Regiment" at "Sheffield". (Presumably he's home on leave). There is no placename that I can see in the original writing, so where "Sheffield" comes from I've no idea. I can see that something lay under "2nd Battalion, Cheshire Regiment" - but I can't see what...

It's slightly disturbing if you start thinking about it...

[Guy]

There were a lot of reorganisations of the Cheshire Regiment during the interwar period with regiments being disbanded and amalgamated etc. Perhaps he was temporarily stationed in Sheffield or even possibly (grasping at straws) in transit on HMS Sheffield (a light cruiser) though I have no idea when that HMS Sheffield was launched.

Cheers
Guy
PS not the destroyer HMS Sheffield

[AdrianBruce]

Guy - the NLS Army Lists don't have much coverage of the inter-war years, so they're no help. However, FMP Newspapers have "Private Amos Capper, of the 2nd Cheshire Regiment, stationed at Hillsbro' Barracks. Sheffield" in the 25 June 1921 issue of the Sheffield Independent, so it looks like they were there at the appropriate time. Still, it's not clear to me whether the enumerator copied something now no longer visible or whether he simply "knew" where the 2nd Battalion was, maybe from other schedules...

[woodchal]

Maybe cynical view - BUT When you are selling transcriptions and images by the record a little bit of inaccuracy can help drive sales.

[Guy]

Not only is you suggestion wrong it is insulting!

You obviously know nothing about the task involved and the facts behind the transcription.
The National Archives put restrictions on the way the transcriptions were done which in itself reduced the accuracy of the transcriptions. Instead of the transcribers seeing a full page they were limited to sections of the page. I do not know at the moment whether those sections were in columns as the 1939 or were horizontal sections.
The problem caused by this is the transcriber cannot search the page for a clearer example of the word as would happen in other transcriptions, depending on how the sections are oriented the transcriber cannot even check that a sentence makes sense.
An additional difficulty was added due to covid, this put restraints on the number of people working on the transcripts and the amount of checking that could be done.

As someone who has done a lot of transcriptions in the past and understands how difficult it is at the best of times, I think FindMyPast have done an excellent job and should be congratulated.
Cheers
Guy

[Thunder]

I disagree with Guy on this. I doubt TNA put restrictions on how the transcriptions were done, unless you know where this information is and I await a response from the Information Commissioner's Office on the issue. The columns should have been clear, for example the entry for Winston Churchill (transcribed as 'Churchil') has his place of birth transcribed as 'Blantyre' (which is in Scotland!) whereas the image clearly says (correctly) 'Blenheim'. As far as checking the transcriptions go a lot of the errors would not be picked up and TNA did award the contract three months after the date the contract was supposed to start!.