The 1939 Register – when is a ‘census’ not a census?

Detailed instructions for enumerators covered every eventuality (catalogue reference: RG 28 211)

Detailed instructions for enumerators covered every eventuality (catalogue reference: RG 28/211)

The declaration of war on Germany in September 1939 came as no great surprise. Preparations had been made well in advance at national and local government level, and homes, businesses and institutions had made their own contingency plans.

Troops were mobilised, the blackout enforced, and schoolchildren evacuated from cities. Considering that this was at a time of national emergency, things went remarkably smoothly on the whole. This was no coincidence, because in many cases these plans were based on experience gained during the First World War. This was particularly true in the case of the National Register, the most comprehensive survey ever taken of the British population, carried out on 29 September 1939.

If there had been no war, or a short war, the census would have been taken as planned in 1941. In fact, most of the administrative framework already in place for the 1941 census was used to compile the National Register, which made very good sense. In England and Wales the census was due to be organised as before, by the General Register Office and local registrars, who recruited enumerators to go door to door. The difference this time was that they were recruited two years in advance, so that they could be deployed immediately in the event of war. There was also a second tier of enumerators at the ready, only to be called on if a National Register was taken, and not required for a census in 1941.

These extra men and women were needed because the National Register needed to be taken very quickly and accurately, so the enumeration districts were divided into smaller units that one person could cover over the allotted weekend. So far, the National Register looks a lot like a census, but there are important differences:

‘A census is solely directed to the compilation of statistics… on a given day. It is concerned with any person not as an individual but as a unit … [and] is not concerned with any particular person when he has once been reckoned in his appropriate category. The fact, therefore, that the census returns are very largely incorrect on the day after that upon which they are made does not appreciably invalidate the statistics.’ (catalogue reference RG 28/1)

This is taken from a memorandum written by Sylvanus Percival Vivian of the National Insurance Commission, shortly before the introduction of a National Register – in 1915! This was supposedly a means of more efficiently organising the civilian work force, but as many people suspected at the time, it would also be a very useful recruiting tool for the armed forces. This proved to be the case when conscription was introduced in 1916.

Mr Vivian went on to express his concerns about the flaws in the proposed register, this time on the difficulties of updating it once the initial survey was complete.

‘But in the case of a register of population the [purpose is]… to secure information at any time as to the actual personnel of any given district. Changes, therefore, in the residence and status of individuals… are of vital importance to a register, and any scheme for the organisation of a register must necessarily provide adequate machinery for recording… its actual personnel, which is, of course, in a constant state of flux. Registration differs, therefore, from census-taking in the important respects –

a) That it is concerned with individuals instead of statistics

b) That it must provide information as to the position at any time, instead of a single appointed date.

It is impossible to secure the performance of any new duty by the individual members of the population unless the particular act of fulfilment is closely linked up with the personal self-interest of the individual, or with the action which he would normally take in his own self-interest.’

He pointed out that there seemed to be no mechanism in place that would ensure an accurate register as people moved around, joined the armed forces, or died. This was no doubt based on his own experience working in the comparatively new field of National Insurance; the General Register Office had plenty of experience of census-taking, but keeping a register continually updated was a different business altogether.

A report compiled after the war in 1919, proved him right (catalogue reference RG 28/1). The 1915 register listed all civilian males and females between 15 and 65, excluding prisoners, inmates in asylums or Poor Law institutions, prisoners of war and internees. There was inadequate provision for adding the names of those released from such institutions, arriving from overseas, younger people as they turned 15, or men demobilised from the forces. Nor was there an effective procedure for removing the over-65s, or anyone who had died! In theory, the army and the navy would return a form to the local authority for each man who enlisted, but since they had no ‘self-interest’ in this process, it was a low priority, and the forms were returned slowly, if at all.

An unforeseen problem was the cost of replacing, free of charge, certificates of registration that people had lost. Others were the lack of preparation time allowed, and the overly-complicated system of colour-coded paperwork that ensued. These were largely overcome in 1939 by:

  • issuing everyone with an Identity Card on registration, which they were required to carry at all times
  • using the Register for the supply of ration books – so no registration, no rations.

This provided the self-interest element that Vivian had recommended. A charge was also made for replacing lost Identity Cards, along with numerous other bureaucratic refinements that ensured the National Register was kept up to date.

This was done under the supervision of the Registrar General, a post held in 1939 by none other than Sylvanus Percival Vivian!

The 1939 Register for England and Wales, as taken on 29 September 1939, is due to be released online later this year. You can find out more on our partner site


  1. […] The 1939 Register – when is a ‘census’ not a census? […]

  2. Steven Smyrl says:

    From CIGO News, 22nd September 2010:

    UK 1939 National Register – Latest Update

    The Council of Irish Genealogical Organisations continues with its campaign to obtain full access to data recorded about deceased persons from the United Kingdom’s war time National Register. After high profile campaigns led by renowned professional genealogists Steven Smyrl (CIGO’s executive liaison officer) and Guy Etchells, at the end of last year the National Health Service Information Centre (NHSIC) finally conceded a public right of access to data from the National Register. It introduced its ‘1939 National Register Cost Recovery Service’, but then astounded all by disclosing only the data as recorded on the 29th September 1939 and continuing to withhold subsequent annotations to the register about dates of death and changes of surname. The NHSIC invoked section 44(1) of the Freedom of Information Act 2000, which is an absolute bar based upon the premise that some other piece of existing legislation provides an exemption against disclosure. The NHSIC held that such a bar to disclosure was contained in section 42(4) of the Statistics and Registration Service Act 2007.

    However, CIGO believed otherwise and in June decided to challenge the NHSIC’s policy. We requested disclosure of the date of death from the National Register for a Mr. Theophilus Collins Baldwin who was born in Ramsgate, Kent on the 9th October 1847, which made him born 163 years ago! We were aware that Mr. Baldwin was said to have lived to a very great age. In due course the NHSIC refused our application and in turn we appealed to the UK Information Commissioner. On CIGO’s behalf the Information Commissioner approached the NHSIC and indicated that by the 17th September they must either disclose the requested data or apply the exemption and then await the Commissioner’s adjudication in the form of a Decision Notice.

    At the very last minute, on the 17th, the NHSIC disclosed that Mr. Baldwin died on the 24th January 1948, aged 100! In their letter the NHSIC indicated that “in future any requests for information about date and place of death, where we hold this information, will be part of the 1939 National Register Cost Recovery Service”. Given this, CIGO advises that in future when using the Cost Recovery Service specific mention should always be made by the applicant that they require all data noted in the National Register.

    In communication with the NHSIC the Information Commissioner stated clearly that he was of the opinion that “section 42(4) [of the Statistics and Registration Service Act 2007] is irrelevant when it comes to considering whether disclosure would be prohibited [by that Act]”. Further, he questioned the NHSIC’s interpretation of the term ‘personal data’ indicating that “the Commissioner’s definition of ‘personal data’ requires that information must relate to an identifiable living individual. Information about the deceased does not constitute personal data”. In its response the NHSIC admitted that its reliance upon section 44(1) of the FOIA was wrong and that the requested data should not have been withheld. The Information Commissioner’s Good Practice Team has now been called in to work with the NHSIC to “improve its general request handling”.

    CIGO’s success is yet a further validation of its public access policy regarding the National Register and clearly demonstrates how the Freedom of Information Act 2000 can be successfully utilised to the benefit of genealogists.

    1. Iris dempster says:

      Dose this dose this mean that I can look at my dads info about his past life as I can’t get his place of birth , iv been looking for it for 21 years he is not on the 1911 census Evan thouth he was born that year ,he died 43 years ago he is on the bmd only because I put his info on what we have been told so can I access his personal data as I can’t find his name on any thing else as IV been trying to my on family tree plus genealogist have not been able to help me ether so I would be very grateful if I am able to look at my info. Ijd

  3. Eric T says:

    It is perhaps worth noting that, as the 1939 enumeration was not described as a census, analyses of population changes after the war frequently failed to take its results into account. This was particularly significant in the case of Greater London, where comparing the population levels of 1921 and 1951 showed similar results, hiding the rapid increase between 1921 and 1939 and corresponding fall between 1939 and 1951.

  4. Leonard Bainbridge says:

    When & if 1939 National register becomes available to general public I would like to be made aware it may help me to trace family members who dissapeared some were evacuated parents killed no proper records survive

    1. John Phillips says:

      Please make me aware of the 1939 National Register public availability so I can trace my family

  5. Kaye Cole says:

    Please notify me by email when the 1939 National Register becomes publicly available.

    Thank you

  6. diana hage says:

    I would be interested to know in my specific case as I wasn’t born until October of 1939

  7. Janet Woolgar says:

    I would be very interested in seeing the 1939 register. It would help me a lot with my family research, & most likely get me over a few obstacles.
    Please email me when it is available to the public.
    Thank you.

    1. Dorothy Keen says:

      Has this register been published yet? It may shed some light on the whereabouts of my so far untraceable grandmother, if all had to wear identity cards. How does one go about seeking info from the register?

    2. Nell Brown (Admin) says:

      Hi Dorothy,

      Yes – and you can find out more about the 1939 Register here:



  8. Ellen Frost says:

    The 1939 Register could help me find my relative who went missing after 1950, it could provide me with some essential clues.

    Thank You.

  9. Neil Spurgeon says:

    Please inform me when the 1939 Register is published as it will enable further studies into both family and local history research. Thank you.

  10. Sue Hayter says:

    It will help me find what happened to my grandfather who disappeared circa 1925, but evidence suggests he was still alive.

  11. Joan Stevens says:

    I would be very interested in the 1939 National Register and it would be of great help in my work on my Family History. Please notify me by email when it is released for viewing.

    1. Robert Campbell says:

      Dates of death and changes of surname would greatly assist in discovering what happened to my great aunt after 1939. After the 1939 National Register comes online later this year, will this show a register number from which subsequent annotations will be easily accessable from the NHSIC upon application? If this is the case have that organisation indicated what additional fees will be applicable?

  12. Annette says:

    My mother and her siblings left London as they lived near the Air base at Northolt so would be very interested in this information. My Mums best friend was held in detention one day after school and as a result her and the teacher were killed by a Bomb. 🙁 It might help me locate the school.

    1. AUDREY COLLINS says:

      What a sad story. Did you know that civilian casualties of the Second World War are included in the Commonwealth War Graves Commission website ? You might be able to find the information you want there.

  13. AUDREY COLLINS says:

    A number of people have asked to be notified by email when the 1939 Register is released online. The link at the bottom of the post takes you to the Findmypast 1939 Register page where you can sign up to receive the latest news on the project and, of course, the announcement when it is released

    1. Pat Reynolds says:

      Of course, the data is behind a paywall. Which means although it’s public data, paid for by the public, it isn’t available to the public.

  14. AUDREY COLLINS says:


    If you find your father in the register it will give you his date of birth, but not his place of birth. You may get some clues to his background if he is with other members of his family, though. Good luck.

  15. AUDREY COLLINS says:


    Anyone born after 29 September was included in a separate register called the Birth Register, which is not part of this release. If you had been born a month earlier you would be included in the Register, but your entry would be closed since you were born less than 100 years ago, and you are still with us.

  16. AUDREY COLLINS says:


    Where a change of name is recorded in the Register the person is indexed under both names, and the search result will show all of their names. It will also show their identity card number, which later became their National Health number. You would need to ask the NHSIC what their plans are, if any.

  17. Terry Cook says:

    I would like to be updated re the availability of the 1939 register.

  18. Maureen Rankin says:

    I have been very interested in this 1939 register since you mentioned it some time ago. can you still put other relations on it ?
    cheers Maureen Rankin

  19. Joan Watson says:

    I would like to be updated re availability of 1939 register.

  20. Nicola says:

    Findmypast have just announced that they will be charging almost £7 to view a household record when they release the 1939 Register on Monday.

    Are there any plans to make available at TNA itself in the near future so this ‘public’ record can be viewed without having to pay this exorbitant charge?

    1. Nell Brown (Admin) says:

      Hi Nicola,

      Thanks for your comment.

      The charges will be waived for visitors viewing the 1939 Register in our reading rooms once the Register goes live on 2 November.

      We’ve got information on our website to help with planning a visit: – please note that we are closed on Mondays.



  21. beryl hatton says:

    Why is this information not FREE ON LINE as it is not possible to go down to National Archives. Seems the public are being totally blocked from obtaining what is called FREE information. All known sites keep credit card details if you opt for their advertised FREE services and this is not acceptable. FREEDOM OF INFORMATION should be exactly that and not subject to retaining credit card details.

    1. Nell Brown (Admin) says:

      Hi Beryl,

      We are aware that not everyone is able to visit us in person and we will continue to make as many records as possible available online. Conserving, scanning and transcribing the Register has been a complex and lengthy task, and one which we could not resource without the aid of a commercial company. The register is free to search and the pay per view cost includes maps, photos and newspapers not previously available online. As with all records, the 1939 Register online is available in our reading rooms for free.



  22. Pamela Willing says:

    I would like to ask if The National Archives were aware that Findmypast intended to exclude the 1939 Register from the records covered by their subscription and charge extra to view the records, when they granted FMP exclusive rights to digitise the records? Many people make use of subscriptions held by local libraries and record offices to view FMP records for free but this will not be possible in the case of the 1939 Register and the majority of these people will be in no position to make a trip to Kew. I feel that the public are being held to ransom by this company and I would further ask, if TNA were not aware in advance of this situation, will this affect future decisions when they next deal with FMP?

    1. Nell Brown (Admin) says:

      Hi Pamela,

      The National Archives were aware that Findmypast intended to launch the 1939 Register as a pay per view service. In making this service available online for the first time Findmypast have invested hugely in the cost of conserving nearly 7000 volumes (which had been in constant use up until 1991). Findmypast have also taken on the cost of creating the official digital record, indexing more than 40 million individuals listed in the records as well as providing the technical solution which enables the redaction of closed records from each image as necessary. If such a technically complex project was to be economically viable, The National Archives agreed that at launch, the 1939 Register online service would exist on a purely pay per view basis in order for Findmypast to help to recoup their considerable investment. By way of comparison, when it launched in January 2009, the 1911 census was also a pay per view only service.

      The 1939 Register digitisation project was a huge undertaking – on a significantly different scale and complexity from all previous digitisation projects and the multi-million pound project was completed without any public subsidy.



  23. Pete Run says:

    Will this eventually be available to other sites like like the 1911 census was?

    OK not at public cost, but what about all the other censuses that were shared by everyone, why has FMP got sole rights to a “public” asset, yet another cash cow for Brightsolid (or whatever their name is now). The cost angle does not stack up, or else they would do the same for every collection they get. And there was I thinking they had mended their ways since entering the USA market.

    1. Nell Brown (Admin) says:

      Hi Pete,

      All our co-branded arrangements, called Licensed Internet Associateships (you can read more here: are awarded after a competitive procurement tender process that is open to all genealogy sites. Findmypast won the tender process for the 1939 Register after bids were evaluated on criteria such as quality, security and handling of closed data.

      All of our licences are non-exclusive, and some (such as this) have periods of primacy. If other companies wish to licence the material from The National Archives after this period we are happy to talk to them.



  24. David Matthew says:

    It appears that FMP have decided not to release the names of people (like my Mum who died 28 years ago). It does seem unnecessary to not release names as they could be found from birth indexes which are open, have TNA a view on this. As FMP are a private company they are, of course, not subject to FOI.

    1. Nell Brown (Admin) says:

      Hi David,

      The 1939 Register captured information on 41 million individuals. At launch, in excess of 28 million records are open and searchable. Findmypast are continually working to improve the record set in as many ways as they can, and as such are looking into the possibility of cross referencing death indexes in order to open as many records as possible. At this stage, the only way to ensure 100% accuracy to open is by provision of evidence of death in the form of a death certificate. If Findmypast can determine the level of accuracy can be upheld with the results of that work being carried out, they will change their process.



  25. Tim Bradley says:

    I am very annoyed that yet again I cannot access my fathers records without having to pay for something that should be free, I have been unable to find out anything about my family as every timer I try I am asked to pay, as a pensioner myself I cannot afford to pay out money for the information. It is impossible for me to travel to London to get this information , very disappointed.

    1. Nell Brown (Admin) says:

      Hi Tim,

      We’re sorry you are disappointed with the cost. FindMyPast have fully funded the investment in conserving, digitising, creating the digital record, cataloguing and indexing the 1939 Register, several years before it would ordinarily have been made available. This was a huge undertaking, on a significantly different scale and complexity from all previous digitisation projects and was completed without any public subsidy.



  26. Derrick Golland says:

    Very disappointed with first access to the site. I entered my mother’s maiden name – guess what – a transcription error. Other entries also showed that the data is short on clarity. All those emails about what I would be getting lead to a real let down. After all I seem to need to put more information in the search engine to find people than I actually got back as a result. Strange when I put in accurate DOB, place etc, that I also get people from all over the place.

    1. Nell Brown (Admin) says:

      Hi Derrick,

      We are sorry you have encountered a transcription error – as ever, transcribed information is designed as a finding aid to locate the correct record you wish to view, which by the sounds of it you were able to. FindMyPast are committed to an accuracy level of 98.5% across the whole Register. However, with an index that links to a database containing details of millions of individuals some errors will occur. The search parameters for place should link to individuals listed only in that specific location. The Advanced search features allow more specific searches.



    2. Ben says:

      As with all transcribed data, regardless of which organisation carried out the transcription, it is prone to human errors. If you detect errors on genealogy sites you can correct them. However, there are often errors in the original documents, which cannot be corrected, as the agreement is that data is transcribed as written. Findmypast, which undertook took transcribe the 1939 Register, has now included it in their annual subscription package and no extra costs are levied to access it. They have also stated that they will review, in the not too distant future, the “closed records” and if the incumbents are now dead they will become “open records”. The last Government review was in the early 1990’s. It has been an invaluable resource for me in tracing the fate of my family, as the actual DOB’s given can pinpoint an event rather than the Event Registration Quarter given for BMD’s in censuses which are often very misleading. This is because the event quarter (under which the event is logged) can be many weeks after the actual event date and even put the registration in a new year.

  27. Carol Hull says:

    I too am annoyed that we have to pay for this information and people who live in London can get it free. We all are not able to access the archives in London.

  28. Gay Howes says:

    I agree with all the whinges and gripes above, what an anti-climax … I’m joining all those who are cross about the 1939 register. I tried to access it this morning, used credits, not happy about that. More importantly, most of what I wanted to see, like my siblings, is not available. In our house I can only see 2 people and the register tells me 6 people are living there. As far as I can see all this hullaballoo was for nothing. The entire hype has been misleading and a major anticlmax. As with others commenting here, we pay subscriptions to Find my Past plus other sites and we aren’t covered. Just another way to make money, and they aren’t giving me what I want, what I was promised. Not a happy user to say the least!

    1. Nell Brown (Admin) says:

      Hi Gay,

      We’re sorry you are disappointed. Findmypast have never promised that the 1939 Register would be included in their subscription packages and the pay per view enables them to recover their investment to bring the Register online. If the people in the house you are looking at were born less than 100 years ago, they are deemed to still be alive (unless proven otherwise) and their records are closed or redacted for data protection reasons.



  29. Edmund Hulme says:

    The register is a huge disappointment. My father and his brother records are not listed. Also my mother and her mother and father are also not listed. Also, the house in Liverpool where I grew up is also not listed, yet the houses on either side of it are listed. Not a very acurate record system. I actually know the addresses that my family were living at when this register took place so they should have appeared.

  30. Angela Harrison says:

    well I do agree we were mislead and let down in the long leadup to the register being released.

    Its cost me 52 pound (about 120 aussie dollars) to find the records I needed the most, and thats just a start..

    And like others on here its quite frustrating to see (for eg) my dads family home where 8 out of 10 people are deceased, most between 1960 and 1990. Yet only 3 people are visible. And so it goes with all the other records, households full of long dead people whose records are redacted. And now that I’ve paid to view the record, I have to jump through hoops and in many cases shell out more money if I want to unlock their data.

    On the other hand… the records I have viewed have been crucial for my research. They prove irish baptisms from the 1850s and 60s (pre civil reg) are the right people…which proves a LOT of other stuff (parents, family farms, date of birth etc). They have also placed my gt grandad where he died 3 years later…and I was very unsure of the death record for him because its in Yorkshire when all our family are in Lancashire. But this 1939 register record shows him 7 minutes walk from where he died, and has the same DD MMM of birth as a record from Lancashire many years earlier. Proof. Thats the end game, and this register has it. For those reasons, I think a lot of people will forgive the terrible price with time. I did spend 52 pound but specifically I spent about 4 pound per record (in a package) and proven years of work for many ancestors so far..

  31. Gerry Gleeson says:

    I believed that census information was witheld for 100 years due to some people being still alive . Now that the 1939 register has been released and as this is effectively the same sort of information from a much more recent date ( 1939) – Will the 1921 and 1931 Census information be available to view shortly.

    1. nigel says:

      Sadly the 1931 census will never be released as apparently it was all destroyed in a fire.
      We can only hope that the 1921 census might possibly be released early?

    2. Nell Brown (Admin) says:

      Hi Gerry,

      The information in the Register is similar to that found in census records, but was collected for a very different purpose. The 1921 Census will not be available until January 2022 at the earliest as it is subject to the Census Act 1920, while another commenter is correct in stating that the 1931 Census was destroyed by fire during the Second World War.



  32. Alison salter says:

    Also disappointed with the 1939 register release. In accurate from what I can tell as relatives that have passed away omitted from viewing. With regard to Gerry Gleeson comment. Would be interested to know regarding 1921 census also as I know the 1931 census was destroyed in a fire during ww2.

  33. H J Hill says:

    I am interested in the annotations that were added to the 1939 Register in later years (through to the 1950s from the looks of it).

    There is an entry for a woman who I have reason to believe did not registered her true name, nor her actual date of birth; and this has the annotation “DMR 11/8/50”.

    Can anyone enlighten me about this annotation?

    1. Nell Brown (Admin) says:


      The alphanumeric codes in the postings column on the right-hand page were added by health professionals at a later date – we know that in some cases codes have either spilled over from the right-hand page or have been entered onto the opposite page of the record. As this column was not part of the accessioned digital record, and not intended for public access, we do not have access to either the data or what the codes mean, other than the Death (D) code. This will be used by Findmypast to open records of individuals who are no longer alive.

      To submit a Freedom of Information request relating to the ‘Postings’ column data please contact:



  34. Martin Short says:

    I searched for my great grandfather and found him on an address not known to my mother, so on that point I am very happy as this gave me his exact date of birth not known to me, Im unable to find him in the BMDs record. The record has now thrown up another question. His record has a red line drawn through it which a note saying see page 16, can anyone tell me what this means and can I view this page 16?

    1. Nell Brown (Admin) says:

      Hi Martin,

      The 1939 Register was used by the NHS after 1948 and changes of names were recorded in some cases, so added names only relate to those originally in the Register. No entries for people born after the registration date were added, so the Register only lists those alive on 29 September 1939. Crossing through was also used as a form of highlighting a record, often (but not always) in cases where continuation entries have been added on subsequent pages. If the record is continued within the same registration booklet that image should be linked to the subsequent image on the online service.



    2. Martin Short says:

      Thank you for replying to my question, unfortunately in my case there are no continuations only a note to say he was a Captain regular army see page 16. There could be a case that he had a different surname when he was born. Would this be why the cross through has been made?


  35. Andrew P says:

    I’m looking at entries for Liverpool and have noticed something strange. In green biro on the original paper various people have made notes that appear to be married names in at least one case for someone who was 3 years old. The exact note says “Ashcroft” (the married name) CR282 WAA 1239.61. I’m reading this as a note made in 1961 by someone with initials WAA (?). In addition I know this person to be deceased yet they have not been censored but were born in 1936.

    Two obvious questions
    – who (if my interpretation is correct) was making notes on the 1939 register in 1962 and why? I can see from Pamela’s note above that the register was in use up to 1991 but haven’t yet found what that use was and if relevant to this question.

    – how did the team deciding who to censor know that this individual was deceased (they’d be 80 now so less than 100)? Have they looked up every data of birth against the GRO index?

    1. Nell Brown (Admin) says:

      Hi Andrew,

      My comment immediately above should answer your first question. As for the second, the records were used up until 1991 and we know that Death (D) codes were entered into relevant records in the postings column. Findmypast have recorded all of these D codes in order to open more records of deceased individuals.



  36. Gerry Gleeson says:

    Help Please – as I am trying to understand – People who have deceased are not available to view or people still living cannot be viewed within the 1939 Register?

    1. Phil says:

      This info was taken form the FMP FAQ page ….

      “Individuals’ records remain closed for 100 years from their date of birth (100 year rule). Records remain closed for people born less than 100 years ago until proof of death is verified. As more records become open as a result of the 100 year rule, more records will become available to search and view online at Findmypast. Of the 41 million records, approximately 13 million (around 32%) will be closed at publication. The 1939 Register will be updated regularly.

      If you notice an officially closed record in the 1939 Register that you believe should be open, you may be able to request it be opened. Click here for more information. “

    2. Nick Serpell says:

      If you can prove that someone who is locked is actually dead there is a facility for you to apply to have that record unlocked. However, you will have to provide a copy of the death certificate

  37. nigel says:

    Like so many other researchers, my own searches in this 1939 register for people born after 1915 are almost universally being returned as “No Results”. I can’t understand why these people are not being found and listed with a CLOSED label?
    I believe that the digitizing and transcribing may have been done using images that had already had all the redactions applied. This would then mean that the records for most people under 100 years old may have not been entered into the searchable database… hence the lack of results!
    I hope someone can prove me wrong in this, otherwise it could mean that most of the entries currently obliterated by redaction (and therefore unrecorded) will be effectively closed until 2039!

  38. Nick Serpell says:

    I understand people complaining about the cost, but the fact is that if FMP hadn’t digitised these records they would not have seen the light of day. Someone has to pay and the National Archives, like many other areas of public life, has limited resources. I think people sometimes underestimate the cost of painstakingly digitising and then transcribing 40 million records, it doesn’t come cheap.

    It also seems that people had an exaggerated idea of what these records would contain. This was never going to be a census but it is invaluable that it confirms the existence of people in a period where the previous census was in 1911 and the next is not due for release for another seven years

    I, for one, am delighted these are available.

  39. Anthony Wood says:

    In a reply to a question from Martin Short on red lines drawn through a name and reference to another page you stated – “Crossing through was also used as a form of highlighting a record, often (but not always) in cases where continuation entries have been added on subsequent pages. If the record is continued within the same registration booklet that image should be linked to the subsequent image on the online service.”
    This has happened to me in relation to my late mother. I spoke to FMP today and they flatly refused to reveal or even look at the other page referred to.

  40. SueWilko says:

    I have a question about the 1939 Register. I have just paid to have a look at the records my paternal grandfather and grandmother. My grandfather, Horace Edward Manning, is a total mystery, I have his marriage record to my grandmother and his death certificate but no birth record at all. In the 1939 Register it lists his birth twice – one is 27 Sep 1874 and above it within the lines is 27 Sep 1868. Then there are some initials which I can’t quite make out and then a date 24.8.42. But then there is a highlighted line which says “See Con Sheet”. All this just throws up more questions. How can I find out what this means????? How can I get to see the “Con Sheet”.

    Ref: RG101/0377D/010/8 Letter Code: AOAD

    1. Audrey Collins says:


      If different birthdate is written above the one that was originally listed in the register then it is likely to be a correction. A number of different codes have been used by health professionals to annotate the Register pages, and for the most part we do not know what they represent, although 24.8.42 may indicate the date that the correction was made.

      Regarding the ‘See Con Sheet’ annotation, these often include a page number, in which case the page in question can be found using The National Archives reference search, but substituting the new page number for the one in the original RG 101 reference. If no page number is given then there is no simple way of locating the page. However, in all the instances that we have found so far, the entry on the other page has shown exactly the same details for the person as the original entry.

    2. David Underdown says:

      To expand on Audrey’s remark, the page number may be one out from the item number, as the booklets had an unnumbered cover page which was scanned as item 1 if present (some had been lost through damage over the years, and some volumes had been duplicated and rebound without cover pages)

  41. Jackie Cotterill says:

    I assume that the 1939 census will be available at National Archives Kew. Will we have to pay the charges listed or will it be free access.

  42. Myrtle Cooper says:

    My grandfather was born somewhere between 1888 and 1895, I was hoping to get his date of birth from the 1939 register as I know he was still alive then.
    I have found his wife(my grandmother) stated as married, but no grandfather.
    I know they were together in the early/ mid 1930s, and from 1946 onwards, as I use to stay with them every year up until 1954 in the same house where my grandmother is living on the 1939 register.
    I have searched for just the surname(RYAN) in the whole register and variations of it, also his first name and surname, nothing.

    I know he died in 1963.

    My questions are:

    1. How could he get an identity card, ration book, or be on the National Health without being on the 1939 register?

    2. Is there another way I can find his date of birth?

    1. Audrey Collins says:


      Since you did not find your grandfather at home with your grandmother on Registration Night, I am afraid he could be very difficult to find since you have so little other information to go on. There are some helpful suggestions in our own research guide, and some search tips on Findmypast. It is very likely that he is somewhere in the Register, and you may find him yet. However, if he registered late for any reason, or was outside the country at the time, or was a member of the armed forces, he would still have an identity card and a ration book, but his details would appear in a a separate register, which is not part of this release.

    2. Natasha says:

      if he was in the armed forces then he won’t be on the register, I had guessed this as I looked for my grandfather, then today I read a further description which may be the answer to your question

      Armed forces personnel

      The Register was not meant to record members of the armed forces and the records do not feature:

      British Army barracks
      Royal Navy stations
      Royal Air Force stations
      members of the armed forces billeted in homes, including their own homes
      However, the records do include:

      members of the armed forces on leave
      civilians on military bases

  43. Jenny Beckingham says:

    If this was a working document until 1991 then lots of members if my family should appear on the register as they died before 1991, but yet they are still reacted. There seems to be lots of errors on it, for example just one of the errors we have found so far…my husbands grandmother has both her first and last name mispelt.

    Are maiden names before 1939 included on the record or just name changes after 1939?

    1. Audrey Collins says:

      If you find a transcription error you can report this using the ‘Update the record link’ on the full preview page once you have unlocked a record. However, if the information on the original page is wrong, then that will stand.

      Only changes of name after the register was taken will be noted. These are mostly women changing their name on marriage, but will include changes of name for any other reason, such as by deed poll.

  44. Lilian Johnson says:

    I have spent years submitting corrections for transcription errors in Census and BMD listings.
    How can I easily get you to correct this type of error for the 1939 Register?

  45. Myrtle Cooper says:

    Thank you Audrey, how do I find out if he is on the other register that isn’t part of the 1939 register that has just been released, and will that part be released, please?

  46. nigel says:

    I am trying to locate my late father in the 1939 register, but all searches are returning “No Results”. He was born in 1922 and died in 1986, but I suspect that his entry in the register is still “closed”. I can supply the death certificate, but as I have no idea exactly where he was living in 1939, are the TNA / FMP team able to search the “closed” records and locate his entry, so that it can be opened?

  47. Neil Ashton says:

    Were there any groups of people excluded from the 1939 register? E.g. prisoners or those serving overseas?

    1. Audrey Collins says:


      Only people who were in England and Wales on 29 September 1939 are included in the Register. Anyone who was already on active service in the armed forces will not be included, as they were enumerated separately by the military authorities who issued their identity cards. But every civilian had to be registered, including prisoners, inmates of institutions and even vagrants who were sleeping rough.

    2. Natasha says:

      “Armed forces personnel

      The Register was not meant to record members of the armed forces and the records do not feature:

      British Army barracks
      Royal Navy stations
      Royal Air Force stations
      members of the armed forces billeted in homes, including their own homes
      However, the records do include:

      members of the armed forces on leave
      civilians on military bases”

      this was in TNA info found here

  48. Julie davies says:

    I am more confused that before after viewing the household of my great grandfather Charles Percival! I knew he married Esther Thompson (born Goudge in 1876) in 1937. In the 1939 register it shows Charles and Esther living at that address but gives a birth date on 9/6/1894. If the year had been slightly wrong I wouldnt be surprised but 18 years different? Anyone have any ideas

    1. Audrey Collins says:


      It sounds as though you have looked at the actual image for the page, and found that the date of birth is incorrect there. If this is the case, then it is likely that the enumerator made a mistake when he or she copied the information from the household schedule into the enumeration book. In this respect the 1939 Register is very like a census (apart from 1911) where we see the enumerator’s handwriting, but not the original source.

  49. Peter Gardner says:

    From “Expert insights into the 1939 register”:
    One question that has come up frequently is how the transcription process worked. Owing to privacy reasons, rather than dealing with horizontal rows of text each transcriber worked in vertical columns so that they could never see the entire record of an individual whose record would be closed when the Register was published online.

    This would appear to mean that transcribers were more or less working in the dark with no possibility of using both vertical and horizontal consistency checks (i.e. common sense) to achieve the most probably correct transcription of a line of text. Also the process of re-creating rows from columns would seem to introduce a whole new category of problem situations: alignment errors.

  50. Judith LaGrave says:

    Columns not lined up
    The fact that the transcribers worked in vertical columns explains a number of errors I have found. In at least two cases information does not line up correctly.
    I deliberately selected people in my tree with very unusual names. Example 1: husband has birthdate 1915 (which would make him 4 years younger than his son) and he’s described as belonging to a different household. Wife has incorrect birthdate 1876 which IS her husband’s birthdate.
    Example 2: husband has wife’s birthdate; wife (b.1901) closed presumably because she has her elder son’s birthdate (1921); elder son closed; younger son (actually b. 1922) incorrectly open because he has wrong birthdate date (1908. Presumably birthdate of first person in next household.)
    Because of the errors, I have not wanted to waste money unlocking these households but could see from TNA numbers that they were consecutive records.
    Now the TNA records have disappeared from the preview. Is this deliberate FMP?
    I would have liked to report the “problem” pages to FMP as this type of error probably affects the whole handwritten page.
    Also being able to see the TNA numbers does help you to search for the right household

  51. Anthony Wood says:

    There have been no admin replies for a week now. Could I please have a reply to my question dated 3rd November?
    Also can you clarify if on a visit to Kew the original records are available for inspection and, if so, are photographs permitted?
    Thank you.

    1. Audrey Collins says:


      In response to the question immediately below yours of 3 November I wrote:

      ‘Regarding the ‘See Con Sheet’ annotation, these often include a page number, in which case the page in question can be found using The National Archives reference search, but substituting the new page number for the one in the original RG 101 reference. If no page number is given then there is no simple way of locating the page. However, in all the instances that we have found so far, the entry on the other page has shown exactly the same details for the person as the original entry.’ to which David Underdown added:

      ‘To expand on Audrey’s remark, the page number may be one out from the item number, as the booklets had an unnumbered cover page which was scanned as item 1 if present (some had been lost through damage over the years, and some volumes had been duplicated and rebound without cover pages)’

      I hope that this answers your question. Regarding the availability of the Register on site here at The National Archives, an Admin post on 28 October had already confirmed that the Register would be available here, free of charge. This information is also in our 1939 Register research guide. We also encourage the use of cameras (with the flash turned off)to make copies of our records, or you can print out you results for 25p per page.

  52. Peter Faulkner says:

    I just opened the entry for my wife’s great grandmother and her daughter and son-in-law. In the Instructions column on the far right there is the following written in red ink:

    “II see con sheet (Ref ER3B 19… C1…)” The “…” is because the right edge of the image is cut off short.

    Does anyone know what this might mean?

    1. Audrey Collins says:

      The Register was updated manually for about 50 years, first by National Registration staff, and later by health professionals when it became the NHS Register. A number of different codes were used for these annotations, either for the internal administration of National Registration and the NHS, or for the gathering of health statistics. They were usually added to the right-hand page which remains permanently closed, although occasionally they appear in the visible part of the register. However, we do not have a key to these codes.

  53. Harald Wollinger says:

    Unfortunately, I have to agree with a lot of the previous comments. The first use of the database and unlocking of households has not instilled a lot of confidence.

    I searched a number of family members who were born over 100 years ago and have all died. However, neither of them appears in the results. As closed records do not show up even with the basic information, it is not possible to ascertain whether these were simple transcription errors or errors in treating the record as closed.

    Having “unlocked” several records, I fail to appreciate the value for money equation in this process, so will be very careful in using this database going forward.

  54. Judith LaGrave says:

    TNA Reference numbers on Previews
    Queried FMP about why they have removed the TNA numbers from the previews. They replied that this facility was “not being utilised in the correct way”. Not a clue what they mean. But I’m not happy that a useful search tool was removed.

    1. Nell Brown (Admin) says:

      Hi Judith,

      If you believe there to be an error with transcription you are able to report this to FindMyPast within the 1939 Register pages. You can search using TNA references within the advanced search.

      We hope this helps.


    2. EJ says:

      I suspect it means they realised people could use the references to identify households without having to pay. What other reason could there be? Why would it matter how it was being utilised otherwise.

  55. Sue Cocker says:

    Great resource, but at least two roads in Grays, Essex seem to be missing. Who do I contact to find more details? Find My Past flummoxed.

    1. Audrey Collins says:


      Searching for addresses can be problematic in the 1939 Register, as it can be in the census. Detailed and updated addresses were held at local level, and not in the central register books which allotted only a small space for this information. So some addresses will be there, but are hard to find because of the way the addresses were recorded.

      Having said that, there is always the possibility that a page or pages have been missed in a record-set of this size. There is also an outside chance that some household returns were received too late to be included in these register volumes, but were included in a separate register which was used for late registrations, new arrivals into the country, and so on.

      If you have conducted searches by name, and by street, using wild cards etc, and a street still cannot be found despite all your best efforts, you can send an enquiry using the ‘Contact us’ form at the bottom of any page of our website, and we will investigate.

  56. Martin James says:

    When the register was launched the TNA References were available in previews, however since last weekend they have begun to disappear. Despite asking a number of times FMP have not provided any satisfactory answer for their removal.

    The TNA Ref was useful in helping to identify those ancestors with common surnames and avoid wasting expensive credits on speculative record unlocking.

    If TNA references are available on census transcripts and the 1939 Register is a “census subsitute” then why have they been removed.

    At the outset tree must have been a design agreement and if they had an issue with the reference then it should have been raised then not 7 days after release, when the genie is already out of the bottle.

    Can the National Archives overrule FMP on this matter or do you have no control of what FMP do with the records under TNA control.

    1. Nell Brown (Admin) says:

      Hi Martin,

      You can still search using TNA references within advanced search.

      We hope this helps.


    2. Martin James says:

      Sorry Nell not the answer I was looking for, now the references have been removed, makes that search function only available if you open records.

      No explanation has been offered for their removal, if it was OK pre-launch on the first week why have they been removed.

      We deserve an answer or do I need to submit a FOI request?

    3. Nigel Osborne says:

      Nell Brown (Admin) – re your answer to martin:
      “You can still search using TNA references within advanced search. We hope this helps.”

      Can you explain how this can be of any possible use when there is (now) no way to know what the TNA reference is?

      And before you say that I can pay to open the household record – why would I do that until I know it is the correct household? I might have to pay a dozen times before I get the right one (clearly the reason for FMPs change of practice) when searching on the TNA reference would give me all the open names on the household record BEFORE I pay.

    4. Nell Brown (Admin) says:

      Hi Martin and Nigel,

      Findmypast have removed The National Archives catalogue reference from search results in response to concerted efforts by some online groups to share and even build apps and spreadsheets encouraging the manipulation of the embedded file references which allows people to access large amounts of information without paying. The remedial action was taken by Findmypast to prevent this systematic breach of their terms and conditions and fair usage policy and to protect their income.

      You can still search using The National Archives’ catalogue reference in the advanced search and the reference is returned when accessing the full transcript and image.



    5. Nigel Osborne says:

      Sorry Nell but you have completely missed the point in your comment:
      “You can still search using The National Archives’ catalogue reference in the advanced search and the reference is returned when accessing the full transcript and image.”

      If you pay to open a household and access the reference from the full transcript, then there is NO NEED to use the Advanced search on the reference. It just gives you the same information you can see on the image anyway!!!

      So without knowing the reference, that part of the Advanced search is completely pointless and might as well be removed along with the TNA reference already gone.

  57. Nyall Meredith says:

    On the images of the register is an end column headed ‘see instructions’ which has not been transcribed. In one record I’m interested in the notes in this column have been partly obscured by the blanking out of the following closed record.
    Is there any way I can access this ?

  58. Tony Margrave says:

    This sounds like a very important and valuable resource. It’s a pity we have to pay to read it. Family research throw up lots of inconsistencies and sometimes many names have to be checked to get the right one. This can be costly. Also the availability of this resource invites the question, why are the census records from 1921, and 1931 not available? What is the difference between these and the 1939 register? The Americans have made all theirs available down to 1941 I believe and so what are we hiding? I suppose Find my past funded the project and have a monopoly claim over it which is a pity because I would regard Ancestry to be the better research tool and I am sure many of us already pay for that service. I imagine they would have been happy to cut a deal with the NA.

    1. Nell Brown (Admin) says:

      Hi Tony,

      The 1921 Census will not be available until January 2022 at the earliest as it is subject to the Census Act 1920. The 1931 Census was destroyed by fire during the Second World War.

      I hope this helps.



  59. Moira Hill says:

    Found a similar comment in the “see instructions” column of the 1939 Register for my Grandmother as the of the other comments above, mine says “See page 17”. I hadn’t realised that you could click on the arrow to the next page until a friend suggested it. Only problem is that I can’t see in anyway that this person is connected. She hasn’t got the same schedule number as my Grandmother and is no relative. How do we confirm that this is the actual page needed when we can’t see a page number and there are so many transcription mistakes?

  60. Vicki Smith says:

    It seems to me that the price is being “justified” by all the extras that go with the records, unless the photos, newspapers, maps etc show something relevant to my family i’m not interested. It’s a lot of money if you have a big family! I can’t afford to pay at these rates for all of my family “households”, maybe more people would be buying more credits if they were less expensive, after all newspapers for an area would only have to be scanned once and then included in multiple households. Also were not volunteers used? Maybe i’m wrong but i can’t imagine many people doing this!

    1. Antony says:

      The extras are rubbish anyway.

      Who want’s information from the local paper for Edinburgh when my family live in Manchester. Just an excuse to steal your money under false pretences

  61. Nigel Osborne says:

    Can you explain what the correct reference is for this data, for sourcing purposes is?

    I have seen both R39 and RG101 referred to:



    1. Audrey Collins says:


      The references beginning TNA/R39 are image references used by FindmyPast, which contain elements of the actual document reference. A document reference is in the format RG 101/2111C which represents an enumeration district (ED) – in this example the ED bearing the letter code DVIL in Chesham Urban District. A number after this eg RG 101/2111C/24 identifies a page within the ED, and this is the reference you need to locate that page image again, using the ‘TNA Reference’ facility in the Advanced Search ie (Piece Number) 2111C and (Item Number) 24. So in answer to your question, the correct citation for a page image would be in the format RG 101/2111C/24. Most of the piece level catalogue descriptions have now been uploaded to Discovery, our online catalgue.

    2. Nigel Osborne says:

      Thank you Audrey. That’s excellent help.

  62. […] so no registration meant no rations! There is a great article on the National Archives website here which gives some more background […]

  63. Dave Tomlinson says:

    I do not understand the withdrawal of the TNA reference from the Findmypast results.

    If you search using a reference you get a list of the open names on the page. It does not show any of the (very little, but important) extra information available if you open the record. What that list gets you if, like me, you are very familiar with the locality and the likely names of neighbours, is a certainty that you have found a record worth paying for.

    Unless the people who are writing code and using spreadsheets have found a way to access the preview/images etc. (which I doubt) the only thing doing so will achieve is provide customers with more certainty about what they have found and an increased likelihood that they will pay for a record. Thus I believe the decision to remove the reference will be counter-productive.

  64. Tony Proctor says:

    Nell, I’d like to make a small comment on a statement you made above: “We’re sorry you are disappointed with the cost. FindMyPast have fully funded the investment in conserving, digitising, creating the digital record, cataloguing and indexing the 1939 Register, …” (3 Nov 2015).

    I appreciate the investment (cost and effort) that has gone into making this register available. However, higher access costs do not equate with greater revenue from access, and could even have unwanted side-effects as demonstrated in this instance. This applies to any service, and not just subscription costs to access online content.

    The problem is that the majority people have pockets only so deep, and will be unwilling or unable to just keep on paying out at this rate. Much is made of “unlocking your household”; well, I wasn’t born then, but I know most of the information about my direct ancestors. The interesting applications are for the many elusive non-direct ancestors, and for solving brick-wall problems. Unfortunately, that means unlocking, or browsing, many households, and so the practicality of using this great resource diminishes rapidly.

  65. David Sears says:

    The problems highlighted in the discussions so far become obvious within a few moments of using the data, so I guess they will be experienced by most users. The explanations are also helpful and would be worth publishing in advance to manage expectations.

    One unexpected benefit has been to identify dates of birth or even first names for marriages of very common surnames – identifying the wife of someone called Smith would normally be impossible without further info or buying a lot of GRO certificates at nearly £10 a time!

    Unexpected addresses can also be useful – and may confirm memories of sudden departures to the country, assuming that war was imminent. The handwritten notes do yield gems, but only if you are lucky!

  66. ray says:

    As this is an important national data resource, when Findmypast won the contract to digitize the 1939 register were any conditions imposed on them as to the quality of the end result?
    (You may feel a burning second question here- either ‘if not, why not?’ or ‘are you happy that they have complied with these conditions’)

    1. ray says:

      A simple question Neil.

    2. Nell Brown (Admin) says:

      Hi Ray,

      Apologies for the delay in getting back to you.

      All of our licences are commercially confidential, but Findmypast and The National Archives have agreed a set of measures in the form of a Service Level Agreement for monitoring performance under the terms of the licence. Key elements of the service such as access conditions, transcription accuracy and functionality are also regulated by the contract during the term of the agreement.

      I hope that helps.


  67. Jim says:

    The 1939 Register is a superb source of information both social and genealogical that has been completely overpriced. Just who is interested in the additional so-called valuable extras?
    I have a very active interest in both aspects ……..but the price is exorbitant!! For instance I have something approaching 130 households that I am interested in on the genealogical front, but I am not able to afford perhaps 4 or 5 times that number of “unlocks” to find out the information that is so poorly transcribed. At least if the TNA reference numbers were still available I would at least be tempted to pay perhaps 10% of the current rates. On the social history side of my interests I am afraid that the costs have ruled out any thought of using this facility.
    No doubt some changes will be made in the not too distant future, I am sure my experiences are little different to most other would-be users.
    A small token of understanding would be signalled by making the TNA references available again NOW; as for the many other problems these will have to wait for another time.

  68. I Say! says:

    It was interesting to see the pr coming out of FMP before it released the 1939 register. For ages no release date given , no ideas of cost and then bang -out it came.

    I suspect FMP management were desperate to get the product out well before Christmas so that those “desperate” to research paid the highest price.

    As Christmas approaches it would seem likely they might do some kind of offer , and anyway after a time next year (if you consider this is a product and will therefore follow the normal product life cycle) they must drop the price once they have exhausted the % of interested genealogists who will pay the present price.

    As the national archives have said – there is a time limit (how long?) on how long FMP has exclusivity so it must play a cat and mouse game with its prices. If it holds out too long at high prices with a low volume take up it will not maximise its revenues.

    Consider also its costs are all front loaded – it needs to start bringing in the cash to “payback ” that investment

    I have read with interest all the issues and disapoointments users have fed back regarding the register and on that basis (though of course I’m keen to review 400 odd households) I will wait. Price must come down and after this there is nothing new to look forward to until 2022 anyway !!

  69. Michael says:

    This is a question to TNA.

    Notwithstanding the confidentiality aspects of your dealings with FMP over this matter, you must have some documentation detailing the expected customer take-up of paid access to this register and, consequently, detail of the potential income stream for TNA.

    It would be very interesting to see how this expected figure compares with the actual figure.

    Is that information freely available, or would it require a FOI request?

    1. Nell Brown (Admin) says:

      Hi Michael,

      We worked with FMP to ensure revenue and royalty forecasts were achieved. These forecasts are commercially confidential and therefore not freely available; commercially confidential information is covered under FOI exemption section 43 (2).



    2. Michael says:

      Nell. Thank you for your reply. If you are unable to provide information on income, can you please just supply the number details.

      How many accesses were projected and how many have been made?

  70. Samantha says:

    I have been trying to trace my family tree in my spare time. Is there any other way of accessing the 1939 register eg. at the national archives? On the findmypast website they want £6.95 per household – which is quite a lot and its not available via a cheaper monthly subscription.

    1. Nell Brown (Admin) says:

      Hi Samantha,

      The 1939 Register online is available in our reading rooms for free. We’ve got information on our website to help with planning a visit: – please note that we are closed on Mondays.



  71. Michael says:

    Given that only a limited number of people will be able to take up the free access at Kew, what are TNA doing to ensure that the rest of the English and Welsh population is not put to a significant disadvantage?

    What plans do you have to provide this service to people in Wales, or the North, West, or Midlands of England?

    Surely some stragegic computer access points could be provided at County Archives’ offices (for example) so that a national service is provided.

    1. Nell Brown (Admin) says:

      Hi Michael,

      We have a statutory obligation to make public records available on site at Kew; we know it isn’t possible for everyone to visit, and one of the reasons we work with partners to put records online is to make them more accessible to people across the UK, and worldwide.

      In the past, we’ve done something similar to the computer access points you suggest (we paid for a set number of credits for regional hubs for access to the 1911 census), but at this stage we have no plans to do the same for 1939 Register.



    2. Michael says:

      Thank you. Interesting.

      Was the possibility of these “national” hubs actively discussed for the 1939, or was it never even considered?

    3. Nell Brown (Admin) says:

      Hi Michael,

      The possibility of ‘national’ hubs has not been considered at this time.



    4. Michael says:


  72. Michael says:

    Another one for TNA.

    I can find the invitation to tender document, but can’t find the actual contract. I understand that all government contracts have to be made available to the public.

    As the contract appears to have been awarded on 14 February 2014 (source TNA Annual Report 2013-2014), I searched on, but came up with nothing; even extending the date to the end of the month gave no results even though contracts are listed on there from 2012.

    Can you please advise where the published contract details can be accessed.

    1. Nell Brown (Admin) says:

      Hi Michael,

      This is not a procurement contract – it is a licensing agreement. We do not publish those as they are commercially confidential.



    2. Michael says:

      OK, fair enough. Thanks.

      Can you comment then on the estimated value specified on the tender. It says 20 000 000 GBP (otherwise £20 million UK).

      Is this the estimated value for the initial 5 years, or for the initial 5 years plus the option for the full 4 further periods of 5 years each – making 25 years?

    3. Nell Brown (Admin) says:

      Hi Michael,

      The estimate of £20m was a notional figure of what the collection might be worth to a publisher over ten years. This was an estimate two and half years ago, and may not bear any realistic relation to the actual figures.

      The period of the licence is ten years with an automatic roll over every five years if both parties are agreeable.

      I hope that helps.



    4. Michael says:

      Thanks for that.

  73. Michael says:

    Can this question also be answered please from earlier – appears to have been missed.

    If you are unable to provide information on income, can you please just supply the number details.

    How many accesses were projected and how many have been made?

    To clarify this, how many individual records have been unlocked, either paid or free or both.

    1. Nell Brown (Admin) says:

      Hi Michael,

      Apologies for the delay, but you’ll have to bear with us for a little longer on these. Your questions haven’t been missed – but some things take longer to find information about and answer.



  74. Helen says:

    I have had a good result in locating information about my grandmother who I had no information on after 1920. However her information includes her new surname after remarrying in 1946 plus the code DPL. Do you have any idea what this would stand for?

  75. Barry Fisher says:

    I also think the cost is high and should be reduced.We know it has cost FMP a lot of money but as they applied to get the licence they would know it would be a lucrative business in the long run.How much did FMP pay the TNA to get the licence?.
    FMP have no doubt made a lot of money from previous and ongoing records.
    I also cannot find my father who was born in 1913 and died in 1991.
    I would be willing to search the records at a reasonable cost.

  76. David Davies says:

    I have been looking at the 1939 register and have found it fascinating. I have discovered that my mother (then aged 14) and family who were from Gibraltar were in England on day of the register staying with relatives. Astonishingly they went back to Gibraltar soon after( I assume by Boat) only for the Women and Children to be evacuated back to UK following year as War intensified. Second astonishing thing is that (like another comment above) some one has at a later date amended my mothers record overwriting it with her married name. This a bit odd as she did not get married till 1949 some years after the war ended. The records were still clearly being used and updated well after the war. Does any one know why this would be.

    1. Audrey Collins says:

      The records were in use and updated until 1991. Plese see our Research Guide for more information

  77. Nigel says:

    I see on a number of register pages some annotations in red to the right of entries that refer to other pages. The sort of thing I’m referring to says “see page 2″, See page 24” etc.

    What pages are these referring to, and are they accessible to us?

    One of my family has a red line right through her entry, then the annotation “See page 24”. I’d love to know what that is about.

    1. Audrey Collins says:


      These are continuation entries where an extra line was created for an individual, usually at the back of the enumeration book. They usually contain the same details as the original entry; the ‘postings’ information on the right-hand page remains closed.

  78. Brian Arthur says:

    At the end of January I purchased a special one month subscription to FMP for £1, which I saw on Facebook. I was hoping to access the 1939 Register. I did a preliminary search of the register and noted that I would be able to ‘unlock’ the records after 16 February. I logged on and having found my parents record, I found that I will have to pay a further fee to ‘unlock’ their record. On studying your website in more detail, I see that the 1939 Register can only be accessed by those with a 12 MONTH subscription, whereas mine is only for 1 MONTH. No doubt this was mentioned when I took out my one month subscription but I didn’t notice it at the time (small print no doubt). Needless to say I feel very disappointed and a little angry. I feel as though I’ve been conned. I already pay £120 a year to belong to Ancestry and as I am a pensioner, I certainly do not intend to pay an additional subscription just to access the 1939 Register.

    I complained to the FMP Support Team and received the following reply:-

    “Thank you for your email.

    I am sorry to hear that you have been disappointed with the site and hope that this will not deter you from using it again in the future. We do offer a pay-as-you-go option for those who are unable to afford the more extensive subscriptions.

    Having customer feedback is very important to us and this feedback will be used to assist us with future site development. Any suggested site changes have already been read and noted. We shall be reviewing these and where possible suggested changes will be included.

    Please get back in touch if you have any further queries or suggestions that you may have.”

    From what I’ve read on this blog, it’s safe to say that I’ll be sticking with Ancestry from now on and shall remember in future to steer clear of special offers from genealogy websites.

  79. Andrea says:

    On my grandfather’s entry in Derby there is an annotation on the adjoining page ‘other information’ of ‘PWC 238/124’ – it is written in a different pen so presumably entered at a different time. Other entries in this column include ARP, Auxiliary Fire Service. Is there any way of finding out what this means please?

  80. Audrey Collins says:


    There are many abbreviations and annotations in the Register, added by different people for different purposes over a period sixty years. I am afraid there is no master list to these abbreviations so we cannot say with any authority what is the meaning of any of them.

  81. Robert Campbell says:

    Through I have found my great Aunt who went missing after 1939, no death or marriage registration indexed right through to 2007? She was born in May 1895.
    There is a feint pencil annotation what appears as KNG 16-11-44 which of course I do not know what it means. According to your earlier replies the National Archives do not have the key for these alpha/numerical codes, is this still the case?
    Would the subsequent authorities responsible for maintaining this initial registration record after September 1939 possess the key to these codes, in this I refer to the registration authority (local council) during the war years or it’s successor the National Insurance? As she was born 121 year ago can I request additional right hand page information under FOI laws?

    1. Audrey Collins says:

      Robert, I can confirm that neither The National Archives nor Findmypast has access to a key to these codes or their meanings. Many annotations appear in the format that you describe, where the 3-letter sequence may be an area code, and the number is likely to be a date. However, we are unable to confirm that this is the case.

      This register was maintained by the Central National Registration service, and not by the local offices. It was later used as the National Health Service Central Register, transferred to the Health and Social Care Information Centre (HSCIC) in 2007. It is not connected with National Insurance. To submit a Freedom of Information request relating to the ‘Postings’ column data you will need to contact the HSCIC

  82. Doreen Taylor says:

    Transcription Error
    Our Family name is Saunderson so why is it SAUNDERS on the transcript also my Mothers name was Laurie not LOUISE. All this means when you put in their name it cannot be found
    yet you put in the name of street and ten you find the mistakes. I have ask for this to be amended but no such luck. Very Very upsetting.

    1. Audrey Collins says:

      If the transcription does not match the information on the original page, you can ask for it to be corrected using the ‘Report an error’ button. Corrections do not appear instantly, because someone has to look at the entry and agree that the transcription is wrong before that can happen. Sometimes the transcription will not be changed because the mistake is in the original record – often because an enumerator made a mistake, or the information supplied by the householder was incorrect.

  83. Jenny deverell says:

    I have a relative who is redacted in the register who died overseas in the RAFVR in 1942.
    He appears on the Runneymede memorial and the GRO “RAF War Deaths” index.
    Do you still require a death certificate to open his record ?
    Surely his record should be automatically opened.

    1. Audrey Collins says:


      The only records automatically opened are for people born more than 100 years ago, or whose death was notified to the National Register. When a person joined the armed forces they were issued with a new Identity Card by the military authorities, their record was removed from the ‘live’ file of their Local National Registration Office, and no further updates were sent to the National Register. So if a man died in service, this would not be recorded against his entry in the 1939 Register.

      To open a closed record we normally require proof of death in the form of a death certificate. However, in the case of a war death, the information recorded by the Commonwealth War Graves Commission may be sufficient to identify the deceased as an individual in the 1939 Register. However, we cannot guarantee that this will always be the case, so we will deal with such applications on a case by case basis.

  84. Robert Campbell says:

    with reference to your helpful reply on the 14th of March. Regarding the annotation coded on my Great Aunt’s original entry in the Edmonton, London Area. The letters “K N G” indicated a change of address, change of birth year and 16-11-44 indicated when this record was amended. “K N G” indicated the district of Holmfirth in West Yorkshire, where she eventually passed away.
    I beg to differ with you as this information came to me direct from the National Archives itself following a subsequent online e-mail enquiry. So it seems the National Archives DO possess the key to the alpha code as it relates to all the towns and districts. This information is not available to Findmypast to assist their subscribers as all I received from them was and quote “I haven’t got a clue!”. Would it be a good suggestion for the NA to provide a list of 3 letter codes with their corresponding districts to assist future ever increasing researchers using the 1939 register?

    1. Audrey Collins says:


      I’m sorry if I didn’t make this clearer in my original reply. We supplied an explanation of the enumeration districts and a list of 3-letter Area Codes to Findmypast, which has been available on their site since November 2015.

      These area codes form part of the 4-letter Enumeration District codes, which appear at the top of each page, and in the transcriptions. You will also find the Enumeration District codes listed in Discovery, our online catalogue, in record series RG 101.

      However there are many other abbreviations, in a variety of formats, which sometimes appear in the Register, and to which we have no key. These include some 3-letter abbreviations which are NOT area codes, hence the caution expressed in my previous reply!

  85. Vanessa Wright says:

    I was absolutely shocked to find that I have to pay $240AU to access FMP records. I only wish to look up records on the 1939 register as I am already paying a small fortune to Ancestry to access their records. I feel these records should not be the monopoly of one research site but shared with all the major family research sites.

    Is there any possible way of having a system where we can access just the register as I don’t want all of your world wide records. And don’t tell me Credit system either as that is far more expensive and very limited to the number of households that can be accessed.

  86. Antony Lambert says:

    Burntwood Asylum Lichfield Staffordshire (commonly called st Matthews)

    I cannot find this anywhere on the 1939 Index. What’s the code please?

    1. Nell Brown (Admin) says:

      Hi Antony,

      Thanks for your comment.

      We’re unable to help with research requests on the blog, but if you go to our contact us page: you’ll see how to get in touch with our record experts via phone, email or live chat.

      I hope that helps.


  87. Kate Rimmer says:

    I would like to know how you monitor the number of records of living people under 100 years old which are opened?
    Earlier this month (Nov 2016) Findmypast announced that 2 million more records had been opened on the 1939 Register, and a lot of people have said that they have found that the records of some of their living relatives have been opened in this update. Some of us have suggested to Findmypast that they should roll back the latest update until they have found what went wrong, to protect living people’s details, but all they have done is to ask for details of the particular records which individual researchers have found, so that they can close them.
    To give a rough idea of the scale of the problem with the 2 million 1939 Register records opened this month, I took a sample of 100 of the “Living Life Peers” on Wikipedia’s list, born between 1917 and 1938, excluding Scottish and Irish peers, and I managed to find 4 of them on the 1939 Register, all marked as “New” (i.e. opened in the last 30 days), and with enough information on the image to confirm that they are the same people as on Wikipedia. This suggests that there is a big problem with living people’s records being opened, and I would have thought that Findmypast were breaking the terms of their licence for this database by opening the records of maybe something like 140,000 living people (very rough estimate). Of course the terms of their licence are not publicly available for us to check this. But since they do not seem to be in a hurry to remedy this mistake, perhaps The National Archives should take action to make them fix it.
    By the very nature of this database, the people whose privacy has been infringed are in their late seventies, eighties or nineties, and are unlikely to have an annual subscription to Findmypast, so many of them will be unaware that their privacy has been breached but this does not mean that everyone should ignore the issue!

    1. Carianne Whitworth says:

      Hi Kate,

      The National Archives takes the issue of data protection of living individuals very seriously. Findmypast have been working to open as many records as possible under the strict guidelines set out by The National Archives whereby they can use death data to attempt to match records listed in the 1939 Register. The stipulation is that the death data matching has to produce 100% unique matches in order for Findmypast to open the records. For example if two individuals have the same name and date of birth in the 1939 Register which match with a record in the death data Findmypast cannot be sure which record to open so they will not open either, hence the 100% unique match. We are confident that Findmypast are adhering to this methodology. On this particular issue which you have raised we are currently discussing and analysing the situation with Findmypast and will provide a more comprehensive update next week.


      Carianne (admin)

    2. Nell Brown (Admin) says:

      Hi Kate,

      As promised, we have conducted a review of the processes used by Findmypast to match death records data with the details of those listed in the 1939 Register. As a result The National Archives are happy that Findmypast are adhering to the methodology we have stipulated. Of the newly opened individual records a figure approaching 260,000 across whole households have been viewed online. Findmypast have received 28 take down requests equating to an error rate of 0.01%.

      While Findmypast are adhering to the methodology whereby the data in the 1939 register has to have a 100% unique match with the death data we accept that there are instances where errors can occur. This can be for a number of reasons which include:

      – errors in the original records
      – transcription errors
      – errors in the death data used
      – misalignment on image

      Findmypast are adhering to The National Archives’ published Take Down policy to remove the item at once and then to effect an investigation.

      In total since launch, Findmypast have received 51 accurate takedown requests. We are continuing to monitor take down requests. We are also continuing to work with Findmypast to refine the process so that they are able to minimise the number of errors further.



  88. Hilary Greenwood says:

    I have found my relatives but their names are crossed out in red with the comment, written in red, see book 2 page 9. I have asked Findmypast what book 2 is, but they do not know.Does anyone here know, please?

  89. Kate Rimmer says:

    Thank you for your reply, Nell, but I do not feel that the situation is satisfactory. People who view a 1939 Register record may not be aware that the person whose record they view is still alive, or even if they are aware, they may not know them well enough to ask for their ID so they can send it in and request that the record be closed. Or they may just not care about it. It is the actual data subjects (i.e. the people whose records you have opened) whose privacy has been breached here, not the researchers who are looking at them. As I have said before, most of these people will not have annual subscriptions to Findmypast and will not know that their personal data is available for people to view, and you should be protecting their privacy.

    Since the 1939 Register was originally launched, there have indeed been a few instances where a living person’s record was opened due to an error on the original record, transcription error, or misalignment on image, but you cannot deny that since the early November update whereby 2 million more records were opened, there have been a lot more people saying that they have found a living person’s record has been opened. Obviously it is not easy for us to know whether this could be due to “errors in the death data used”, but if you are honestly trying to say that Findmypast have been supplied with death records exactly matching Dennis Skinner or Michael Heseltine’s names and dates of birth, then you need to question where they are getting these supposed death records from and put a stop to it as they cannot be genuine death registrations. (We can rule out the other three possible causes in their cases as the information on their original records was correct and was correctly transcribed, with the index entry matching up to the images.) I mention these two names as their records were opened in the November update but when I posted them on Facebook, they were very soon closed by FMP. Funnily enough, the record of the other famous person whose name I posted is still open, presumably because he is not a politician so they aren’t worried he will make trouble for them.

    The way in which you are monitoring accuracy of records being opened appears to be flawed, as you are basically saying, “Not many people are complaining about their records being opened”. What you should be doing is using a sample of details of living people who were born between 1917 and 1938, and lived in England or Wales in 1939, preferably people who consent for you to do this, look them up, and see how many opened records you find. If you do this straight after an update, you can see if the opened records are marked as “new” which will show if they were opened in that update. This will give you a good idea of the scale of the problem. I am honestly shocked that you are not taking immediate action to close all the records which were opened in November. People do need to be informed that their records may be open, so they can check and request them to be closed if you are not going to do anything about it otherwise.

  90. Kate Rimmer says:

    I just felt I should say this explicitly although it has been implied in everything which I have posted: by far the most likely reason for the large number of living people’s records which were opened in early November is an error in Findmypast’s software.

    1. Margaret Hirst says:

      Relating to the incorrectly opened records in this last batch of 1939 updates. I reported 2 from my family and Kate Rimmer has found 4 from the Peerage listing. A post on Findmypast’s Facebook page had another 6 (including a familial one from Kate Rimmer), so that’s 12. Then there will be those people who didn’t bother reporting them for one reason or another, those who reported them other than by Facebook, and then those who know nothing about them. (I seem to remember hearing that it was a period of hectic activity for FMP’s Customer Service Department.) One of mine was immediately dealt with, but the second took a bit more persuasion. There must have been a significant problem for the decision to have been made to close the records without any supporting documentary evidence, so I do wonder what exactly this figure you quote refers to.

      On that note, I think it’s very dangerous, even flippant, to produce any sort of figure to try to explain away this significant error. ANY incorrectly opened record is serious and in breach of Data Protection.

  91. Len Hutton says:

    Prior to posting, did you ask a statistician for thoughts regarding Findmypast’s claimed “error rate”?

    In my view the claim is meaningless and it’s very dangerous for you to rely on it as evidence of adherence to methodology.

    1. If I look for say Sidney Hutton born 1911 and view that record, it also shows around 30 other people. I am only interested in Sidney and family so if there was a wrongly opened person on that page I would not notice yet it would account as a record view but no take down?

    2. It assumes the researcher will always submit a take down when finding an incorrectly opened record. That’s a very dangerous assumption to make and cannot be based on any evidence.

    3. As Kate has said the only way to check it properly is to take a sizeable data sample and investigate it thoroughly for errors.

  92. Teresa Goatham says:

    In addition to the other replys regarding the ridiculous statistics provided by FMP, has anyone at TNA bothered to look at how open entries are to be reported?
    The impression given that FMP prefer to err on the safe side in the comment that if there are two people with the same name and date of birth FMP will keep both closed rather than open one erroneously, they actually make it difficult to close a wrongly opened entry.
    I have just found a relation open who I received a Christmas card from a few days ago. Should I close the entry? I looked to see I needed to do. I would have to provide a scan of a current passport, driving licence or national identity card. (Or something else with signature and date of birth).
    Those in the register will be aged at least 77. Many of this age and above do still drive and travel abroad, but many do not. They may well have neither of the first two, and since plans for a national id card were scrapped I don’t know what is meant by the third. I don’t know if the relation I have found has any of this, but if he does how would I get the information? Well, I could ask him to scan it and send it to me. (In this case he probably could, but of course plenty in this age bracket couldn’t, and we live at other ends of the country). I haven’t seen him for some years; would it stress him, to tell him information about him has been revealed publicly that should be private? I could contact one of his children – but I don’t have contact details and they have a common surname. So what am I going to do? – nothing. And most people will probably do the same.
    This is the first time I have seen someone I know to be living open – but that’s quite likely because mostly I’ve been looking at entries for people I wouldn’t know about. Mostly the 1939 adds very little to what I know – as was the case with the above wrongly opened entry. I only looked at it after reading the comments here, which made me realise how I’d been concentrating on cousins with whom contact had been lost; I wanted to know what had become of them. ** I think this is a common reason for research in relatively recent records – so it’s hardly surprising people don’t report errors ** This wrongly opened entry, according to FMP a one in 10 000 event, was the very first one I knew about who I looked for.
    TNA, if you really take the issue of data protection of living individuals very seriously, as you claim, you would think a bit more about FMPs statistics, not post such a poor response, and ensure it was easy to close wrongly opened entries.

    1. Margaret Hirst says:

      Teresa. Contact Findmypast Customer support on FirstSupport [at] findmypast [dot] co [dot] uk

      Explain the situation as you’ve described here, and then await their answer. It will either be, “of course we will close it with no argument”, or “you have to provide evidence”.

      If it’s the second, send a reply to them and ask that it be passed to the Head of Customer Services. He can then make a judgement call and hopefully it will fall in your favour.

      As I stated above, I had one of each response to my cousin and aunt, both of whose records were opened in error. But my aunt’s was closed without further argument once I escalated it

  93. David Matthew says:

    I too have doubts over the claims over the 5 million new entries, perhaps we should be told how many entries are still closed, given that everyone who was born in almost the end of 1916 should be opened by now. TNA have told me that the entry has to match exactly any death certificate. I think most people are aware of the limitations of the death indexes and there are certainly cases where people have the same name and for example what if the person moves out of England and Wales?. If the age is checked then there should be no doubt but the issue is that without a name (I can’t see why just a name could not be released as there are electoral registers which contain equivalent information) but why do TNA have to charge so much for opening the record, I know they will say that it is what Treasury ask for recovery of costs but it still seems a lot to me. In my view I do find it amazing that a private company can retain closed Government records.

  94. Nell Brown (Admin) says:

    Hi everyone,

    Thanks for all your comments; the following response is from National Archives Data Protection Officer, Linda Stewart:

    ‘The National Archives takes the concerns raised in these recent posts very seriously and has been investigating them with the full cooperation of Findmypast. I would like to share with you the results of the investigation, and to explain in more detail how the National Archives applies the Data Protection Act to the 1939 Register.

    The National Archives is aware that in complying with the Data Protection Code of Practice for Archivists, facilitated by the Information Commissioner, which states that ‘access to archival records of living individuals will be subject to closure periods up to a maximum of 100 years’ , there will be a percentage of unknown individuals who will live longer than 100 years, and this percentage represents the level of risk that we have to accept in carrying out our obligations under the Public Records Act and the Freedom of Information Act to make records transferred to us available to the public.

    The risk for the 1939 Register has been adjudged by the Government Actuary Department, based on the likelihood of someone living in 1939 continuing to be alive at age 100, to be 1.5%. Following discussion and agreement with the Information Commissioner’s Office regarding this dataset it was decided that this was a tolerable level of risk.

    By matching death data to records of individuals listed in the 1939 Register in order to make them available, it is recognised that there is also a level of risk, but by limiting the number of death data sets used to trusted sources that risk is reduced. Findmypast have used the expertise and resources of a third party approved by The National Archives. Their core business is death data matching, for which they use licensed death data from the General Register Office for the years post 2007 to the present. The information contained in this data is full name and date of birth. This third party have also used probate records to further inform their work and to weed out duplicate entries. The third party then followed the 100% unique match methodology determined by The National Archives to produce an exact list of records on the 1939 register which could be opened. They then passed this information back to Findmypast. There is a 98.5% accuracy rating using this method – consistent with The National Archives’ risk appetite.

    The methodology was determined by The National Archives in advance of any of this work taking place, and we are satisfied that Findmypast and the third party have followed the methodology to the letter.

    Once Findmypast received batches back from the third part they carried out a % QA and determined that there were no errors in that random % in the 1939 data which would lead them to suspect any issues. Findmypast also carried out a % QA of the opened material once the records had been opened and found that there were no discrepancies, ruling out any software malfunction during the actual opening process. If Findmypast had suspected that there were issues when they opened the 2 million extra records at the beginning of November they would have rolled back the release and investigated.

    Errors can occur in this process for a variety of reasons:

    – errors in the original 1939 Register data
    – errors in the 1939 Register transcription
    – errors in the original GRO death data
    – errors in the GRO death data transcription

    or a software malfunction, which in this case has been ruled out.

    At launch there were just over 27million records made available on the 1939 register. Using a mixture of the D codes (deaths recorded when the records were still in use) listed in the records, and the 100 year rule for opening a further 3.5 million records have been opened since launch. The further release of 2 million records thanks to death data matching has resulted in a total of 32.8million records being available within the 1939 Register.

    To date there have been a total number of 53 requests for takedown since launch and 30 requests for takedown since the opening up of those further 2 million records in November. This against a backdrop of some 32.8million open records equates to .00015% takedown requests. All takedown requests where there is sufficient evidence to locate the individual concerned are acted upon immediately.

    It is worth noting that at the present time Findmypast are not insisting on proof of identity being provided with any takedown request as originally planned. The process has been simplified: once notified Findmypast will remove the record regardless of proof of identity being provided; we have always been aware that could be problematic for older citizens who may no longer require a driving licence or passport. The revised process has still resulted in just a tiny number of requests, well within any anticipated margin for error and well within The National Archives’ accepted risk.

    Requests for takedowns of identified living individuals should be sent to support [at]

    This has been a once only process – Findmypast will continue to open records based on the 100 year rule but there are no further plans for death data matching.’



    1. Margaret Hirst says:

      Hopefully TNA will have learnt from the 1939 exercise and will be more careful in future about the proposed methodology of transcription and quality of data auditing prior to release.

  95. Kate Rimmer says:

    I have sent the following to Linda Stewart in an email:

    I hope you had a happy Christmas. I thought it would be best to wait until after the holidays to reply.

    Thank you for quoting the figure of 98.5% accuracy in opening records based on matches against the death indexes. It is quite remarkable that this exactly matches TNA’s tolerance level of 1.5% of opened records being those of living people. I am not clear how this accuracy figure was measured, though. And thank you for thereby stating that it is acceptable to TNA for up to 30,000 (1.5% of 2 million) 1939 Register records of living people aged under 100 to be opened to the public without those people’s consent or knowledge.

    You say that FMP carried out “a % QA and determined that there were no errors in that random % in the 1939 data which would lead them to suspect any issues. Findmypast also carried out a % QA of the opened material once the records had been opened and found that there were no discrepancies, ruling out any software malfunction during the actual opening process”. I am not sure what you mean by “a %” – does this mean 1%, i.e. 20,000 records out of the 2 million which were opened? If not then what does it mean exactly, please? I struggle to imagine that staff went through 20,000 opened records manually checking each one against death indexes which they would not normally have access to? (By the way, the probate indexes which you mention will not show date of birth, as far as I am aware, and so should not be used in matching against the 1939 Register to open records of deceased people, and also could not be used to check whether the correct records had been opened. Also, it is misleading to say that records were matched on full name since the 1939 Register does not always include the complete full name.) In any case, this process would only show whether records were being opened in accordance with data matching, not whether the process was actually opening the records of deceased or living people.

    You list possible sources of errors as errors in the original 1939 Register, errors in the 1939 Register transcription, errors in the GRO death data, and errors in the GRO death data transcription. I do not see that 1939 Register transcription errors can be used to excuse records being wrongly opened, as this would effectively be saying to FMP that the worse their transcriptions, the more they can get away with. If there is a high error rate in the GRO death indexes, whether due to errors in the original data or in the transcriptions, then those indexes should not be classed as “trusted sources” and should not be used for this purpose.

    I am glad to hear that FMP are no longer asking for proof of identity in order to close a record, but the vast majority of people (all elderly) whose records have been wrongly opened will not be aware that it has happened and so will still not be in a position to request closure of their record. There is no particular reason to suppose that people who view a living person’s record will know that the person is still living, or will request it to be closed. Bear in mind that the people who campaigned for the 1939 Register to be opened to the public in the first place did so because they believed in freedom of information, not in privacy.
    Since TNA and FMP were so dismissive of my finding 4 open records out of 100 living life peers (and didn’t even ask me for the names which I had found), I decided to do some more statistical studies. Unfortunately by then it was too late to look for the “new” tag on FMP to see whether any record had been opened in early November, or if it was already open before that.

    Firstly I looked up the English and Welsh nonagenarians from the “Noted Nonagenarians & Centenarians” website and found that 2 out of 48 records were open, both very famous so it would be common knowledge if they had died. But of course this was an even smaller sample than my original one.

    Then I found that there was a new edition of “Who’s Who” published at the beginning of 2017, and that their staff continually monitor the press etc to update their information, so entries are updated fairly soon after a death. Each person’s full name and date of birth is given, although birthplace is not often stated. I used their online search to select living people born between 1918 and 1938, and again left out the Scottish and Irish people, and people from other countries. I also left out life peers as they might have been included in my previous study. I looked each person up on the 1939 Register, using the full date of birth, and if I got a match, I checked details such as other names in the household, place of residence, etc, to confirm it was the same person, then searched on the internet to confirm that there was no news of their death. I looked up 200 people and found that 23 of their records were open, i.e. 11.5%. So it does look as though a large number of living people’s records have been wrongly opened, well over 100,000. If you think that 23 is a small number and can be ignored, please show this to a statistician who can explain.

    I am sorry to say that even if you do eventually close all the wrongly opened records of living people, it could well be a case of shutting the stable door after the horse has bolted. I only realised just how much information can be extracted from the 1939 Register without having a subscription, or even logging into an FMP account, after my own subscription expired. You can, for example, find out the exact date of birth for anyone whose record is open, by trying each month number and then each day number, until you get a match (this means a maximum of 43 searches as you do not have to enter both month and day together.) Of course this process would be time-consuming for the average researcher, but not for a data miner, perhaps using bots. They could easily work the other way round, entering different dates of birth and collecting all the names of people born on those dates. If you think it is unlikely that data miners are extracting data from the 1939 Register, note that soon after its original release, FMP removed the TNA Reference numbers from the preview screen giving the reason that it was being used by data miners. And if you play around with a Google search (or, say, Bing, Microsoft’s own search engine), you can see that their search bots have already stored some 1939 Register search results, presumably without specifically trying. So it is perfectly possible that data miners have already used bots to extract large amounts of data from the 1939 Register, including the information of living people. The data which can be collected from a free search does not just include what is shown on the search results screen and / or preview screen as a bot could try various different values for occupation, place of residence, etc and collect the names which come back as matches. The “Lost Cousins” newsletter has recently pointed out that FMP searches can be carried out using the URL to specify search fields which do not even appear on the search screen, as long as those fields have been transcribed, so removing fields from the search screen would make no difference here.

    If you think that the 1939 Register doesn’t contain any sensitive data about people anyway, then consider the fact that date of birth is used as a security question by many banks and other organisations, and remember that the data was considered so sensitive that very stringent conditions were put on the transcription process (each individual transcriber could only see name or date of birth, not both.)

    It is apparent that FMP does not have an effective anti-bot system, unlike many other websites where you have to log in to your account, pay a subscription or buy credits, and / or complete a reCAPTCHA or similar “human verification” test before you can search their databases. Of course when FMP’s databases only contained data about dead people, this was an advantage to them as it would mean that researchers googling for the name of an ancestor might come across an FMP index record and then pay to view the image. But it seems reckless not to have put anti-bot protection on the 1939 Register search, since it was known from the beginning that living people’s records would be included.

    As far as I know, the 1939 Register database is the first where records have been opened, not just using a rolling 100-year rule (which is quite common by now), but also by using the process of matching names and dates of birth against death indexes. As such, it should have been closely monitored all the way, with the involvement of a professional statistician to advise on how to do that. I would have hoped that the Findmypast 1939 Register database would set the standard for similar future projects (e.g. the Scottish and Northern Irish 1939 Registers, school and university records, etc) but I can’t see this being the case now. Hopefully any archives considering such a project in the future will take careful note of what has happened with the 1939 Register on FMP.

    Kate Rimmer.

    1. Margaret Hirst says:

      Excellent piece of statistical research, Kate Rimmer. I would like to hope that there will be a positive response……however, I can visualise a generalised 1.5% quote coming over the horizon and the body of your report not being acknowledged or properly addressed.

    2. Carianne Whitworth (admin) says:

      Dear Kate,

      Please find below a response to your comment, dated 9 January 2017, from our Data Protection Officer Linda Stewart:

      “Thank you for your full response, and for your continuing interest in the 1939 Register. The Register, although in some respects a simple, albeit enormous, list of households, is a uniquely complex dataset, both in the circumstances of its compilation, undertaken in haste on the eve of war, and in its subsequent use for different purposes over fifty years, involving many handwritten annotations and amendments. As a public record it details the information that was originally and subsequently created, with varying degrees of accuracy and legibility.

      We have endeavoured to extract your various points of concern and address them individually:

      – ‘…Thank you for quoting the figure of 98.5% accuracy in opening records based on matches against the death indexes. It is quite remarkable that this exactly matches The National Archives’ tolerance level of 1.5%…’

      This is a figure Findmypast worked towards achieving through progressive improvement throughout the project up until launch. The reason for the exact match with the 98.5% stipulated accuracy is that Findmypast invest a lot of time and resource to meet the very high level of accuracy, and once they have achieved that target they maintain it as that is the contractual requirement. Ideally Findmypast (and indeed all online publishers) would achieve 100% but given the nature of the material and sources of information and the potential for error, 100% accuracy is unachievable, even with a 1 to 1 check, of all 40 million plus records as data that could verify each entry does not exist. For example, in a dataset of such magnitude, there will be a significant number of entries where people have, for whatever reason, supplied the wrong date of birth.

      – ‘…up to 30,000 (1.5% of 2 million) 1939 Register records of living people aged under 100 to be opened…’

      This is based on the fallacious assumption that most or all of these errors would result in records being wrongly opened, when the reverse is more likely. Clerical errors in the Register are likely to be evenly distributed, but post-launch openings needed to be actively triggered, principally by death-matching, where an error in either database would fail to produce a match, thus keeping records closed when they should be opened.

      – ‘…the probate indexes which you mention will not show date of birth, as far as I am aware, and so should not be used in matching against the 1939 Register to open records of deceased people…’

      Since the probate indexes don’t include dates of birth, no records could be opened by checking them against 1939 birthdates. As we stated in our previous answer, probate and other records were used to determine the correct information with which to match the records . Probate indexes are useful because of other information they provide; in the 1939 Register we have used them to identify dates of death where the deceased has a common name, but the address and/or the name of a spouse matches the 1939 information. Although in theory all deaths up to 1991 should have been notified and marked with a D-code, in practice this did not always happen, so this would enable the verification of earlier deaths; some which occurred outside England and Wales.

      – ‘…it is misleading to say that records were matched on full name since the 1939 Register does not always include the complete full name…’

      The Register does not always include middle names, although in some cases they are written in full, often as a later amendment. For the substantial proportion of the population who had no middle name, the name in the Register is the full name. If one record has a full name and the other does not, there will be no match, yet another example of the default position being to err on the side of caution, keeping a record closed when the person is actually deceased. Where Findmypast had exact matches of first name_surnames or first names_middlenames_surnames they matched to those. If there was just a first name or surname these records were discounted.

      – ‘…I do not see that 1939 Register transcription errors can be used to excuse records being wrongly opened, as this would effectively be saying to Findmypast that the worse their transcriptions, the more they can get away with…’

      Sources of errors include transcription errors by Findmypast, but these would be more likely to result in records remaining closed when they should be open, because they would result in a failure to match. The chance of a wrongly transcribed date of birth in the Register matching that of a deceased person with an identical name must be very low. It is equally, if not more, likely that a clerical error made when the records were still in use applied a D-code to the wrong person.

      – ‘…If there is a high error rate in the GRO death indexes, whether due to errors in the original data or in the transcriptions, then those indexes should not be classed as “trusted sources” and should not be used for this purpose…’

      I’m not sure what would be considered a ‘high error rate in the GRO death indexes’ but as far as we are aware no-one has suggested this. The indexes from 1984 onwards were ‘born digital’, and not transcribed, and D-codes were still being applied until 1991, so the great majority of records opened as a result of death-matching will be from the post-1991 digital indexes.

      – ‘…Bear in mind that the people who campaigned for the 1939 Register to be opened to the public in the first place did so because they believed in freedom of information, not in privacy….’

      Believing in freedom of information and believing in privacy are not mutually exclusive, but the motives of campaigners, with respect, are irrelevant, since the task in hand is to ensure that current Information legislation requirements are met.

      – ‘…The National Archives and Findmypast were so dismissive of my finding 4 open records out of 100 living life peers (and didn’t even ask me for the names which I had found)…I decided to do some more statistical studies…’

      In our reply of 23 December we did direct you to Findmypast’s takedown link where you could submit the 4 names you have found (along with the further 23): ‘Requests for takedowns of identified living individuals should be sent to’. I am sorry if you thought our response to you finding 4 open records was dismissive, that was certainly not the case as these 4 pertain to living individuals and I apologise if we gave that impression. It was more a case that 4 as a number was not a cause for concern as it was a very small number given the properties of this dataset. When the dataset runs to tens of millions, a sample of 100 or 200 doesn’t even begin to be representative, as any statistician would confirm. Besides which, the size of the sample is only one factor; it is equally important to select a sample that is free of bias, and achieving that is much harder than it looks. Given the age (77+) of the people in this dataset, the 1.5% risk of opening a living person’s entry is not evenly distributed. According to the latest figures from the Office of National Statistics, people in the professional social class who will make up the bulk of the entries in Who’s Who, are statistically more likely to live beyond 77, whereas people in other social groups are not. Therefore the risk is aggregated into this section, so that although 23 out of 200 appears to be a high percentage, when spread across the dataset as a whole it is not.

      – ‘…Of course this process would be time-consuming for the average researcher, but not for a data miner, perhaps using bots…’

      Specifically with regards to bots and mass data-mining of Findmypast, Findmypast are accredited to the International Standard for Information Security Management and have a regular cycle of penetration and performance tests. Findmypast have had no vulnerabilities that would leave them open to potential attacks. Findmypast also have site-monitoring in place that would alert them to irregular activity on the site that may be caused by bots, for example high volume rapid transactional searches of the site. To add a human element to this, Findmypast produce monthly usage reports for the entire Findmypast site broken down to a granular level which are analysed for activity by each and every dataset. There has been no suggestion at any point that 1939 has been used outside of the forecasted expected behavioural patterns Findmypast expect. Expected behavioural patterns include, for example, a spike around the recent release, or in line with relevant promotions. Findmypast’s greatest asset as a company is their data and they monitor its usage very carefully to ensure it is not at risk.

      – ‘…consider the fact that date of birth is used as a security question by many banks and other organisations…’

      Date of birth is often used as an aid to identity verification, but if anyone wanted to misuse the information relating to living people, they would first have to identify them from among more than 32 million open records, and it is has been adequately shown that this is not an exact science – as you have yourself said: ‘There is no particular reason to suppose that people who view a living person’s record will know that the person is still living’. There are far easier ways of obtaining a person’s date of birth – such as buying their birth certificate. The GRO places some restrictions on the purchase of birth certificates (which don’t have statutory force), but only for births within the last 50 years, which rules out everyone in the 1939 Register, whether alive or death. Then there are published sources such as Who’s Who in which as you have found, each person’s full name and date of birth is given.

      – ‘…If you think that the 1939 Register doesn’t contain any sensitive data about people anyway…’

      Our job is to comply with the Information Commissioner’s requirements, as stated above.

      – ‘…researchers googling for the name of an ancestor might come across an Findmypast index record and then pay to view the image…’

      Google search results also produce links to search results on Ancestry, My Heritage and The National Archives.

      – ‘… when Findmypast’s databases only contained data about dead people…’

      Findmypast’s databases categorically do not only contain data about dead people. Leaving aside the birth and marriage indexes, they have long had Electoral Rolls and Directors of UK Companies as recent as 2014. The Probate Calendars 1858-1959 also contain names of executors, many of whom are likely to be alive now, complete with their relationship to the deceased.

      – ‘…it should have been closely monitored all the way, with the involvement of a professional statistician to advise on how to do that’

      As previously stated The National Archives’ professional statisticians worked with Government Actuary Department to produce the methodology for data matching which Findmypast then complied with to the letter.”



Comments are temporarily closed