Ancestry.com provide a way of applying per-user corrections to census records. However, this only supports changes to a personal name, place of birth, or date of birth. I would like to report a relationship problem but I see no way of doing this, so I am looking for a suggested approach.

The 1861 census page of Cheshire in England (Piece: 2560, Folio: 23, Page: 6) shows a household with two "Wife" entries:

BRADY, Samuel  Head
BRADY, Mary    Wife
BRADY, John    Boarder
BRADY, Selina  Wife
BRADY, George  Son

First-off, the surname is really "Bradley" but the hand-writing is too poor to submit an objective correction to the original (a mechanism which findmypast supports but not Ancestry). I know this because John Bradley (from Belper) married a Selina Shepherd, had a son (George) in Lancashire, and then moved to Stockport.

However, it is pretty obvious that the second wife is that of the boarder, John. Unfortunately, Ancestry treats all these roles as relative to the 'Head' of the household. It therefore thinks Samuel has two wives, and George is a son by one of them. This meant it took me ages to find this record, even accounting for surname errors, because I was looking for a John with a spouse named Selina.

Is there a recommended correction that could be applied, and which would assist other researchers?

  • Interesting and well worth highlighting for the effect on who is married to whom in the index...
    – AdrianB38
    Commented Dec 12, 2012 at 22:43

4 Answers 4


I have encountered similar situations. I don't know if there is anything automated that can be done, but you could leave a comment explaining the error & file a report with Ancestry.com. It won't help the immediate indexing problem, but the more such errors they know about, the more likely they are to fix them.


I talked to an Ancestry.com tech support person about a different issue, and as a follow-on, also asked about corrections to fields that are not editable through the web site. His advice was the send email to [email protected] with your ancestry.com account, the URL of the resource that contains the error, and evidence for why the error exists and what the correction should be.

Anecdotally, they seem to keep track the numbers of reports of problems with their data. When I called about the ship's manifest issue, they told that one other person had already reported the problem, and that my report would be added to it as well. So there is likely to be an advantage to reporting these things to affect Ancestry.com's priorities when deciding what to fix.


What you're asking to do isn't a correction, but an annotation based on information you know from other sources.

Ancestry have correctly transcribed the record:

enter image description here

The census enumerator omitted information by describing her as "wife", rather than "wife of boarder", but changing it at this stage is adding a subjective opinion.

Ancestry does allow surnames to be "corrected". In most cases these are helpful, for spelling errors. Many people also add the maiden names of wives, which again can be helpful if accurate. But I've noticed an increasing number of cases where these corrections are just plain wrong, and it's starting to affect the quality of the search results.

So I don't agree that there should be a way for you to add a correction that alters how the record is indexed. An annotation, fine, but to modify the way data is found based on what a user modifies it to (from a correctly transcribed original) is not a good idea.

  • 1
    I'm not sure I agree with the idea that modifying the index is wrong. Modifying the index is not the same as modifying the source data, which remains untouched in the image. While we might want to keep subjective opinion out of transcripts and indexes, the blunt truth is that transcripts and indexes are infested with subjective opinion. Every time we interpret a long "s" as "s" and not as "f", that's subjective, based on our knowledge of what's probably there. Capital "J" and "T" can be confused - but none of us would index a man as "Tames" or "Tohn". However ...
    – AdrianB38
    Commented Dec 12, 2012 at 22:55
  • However in this case I would be hugely concerned about indexing in any way other than as written. We can be fairly certain Samuel did not have 2 wives - I think that's an objective interpretation of a UK household. However, deciding who George's father is, is another matter - I've seen cases where the same ordering and relationships are used and the equivalent of George is a child of the head of household. Which makes me uncomfortable about the risks involved in assuming Selina is the wife of the chap on the line above. Sure she is - here. But I worry about the other possibilities.
    – AdrianB38
    Commented Dec 12, 2012 at 23:05
  • I agree that a policy of recording the record verbatim would preclude this change. However, Ancestry does claim to support subjective (i.e. per-user) changes, including annotation. Contrast this with findmypast who only support objective changes which are then made to the core transcription. The current data is clearly wrong, and breaks some implicit rules about marriage and spouses, so a dogmatic policy does not help here. In a similar vien, a transcription that includes a missing surname when the enumerator implied it was the same as above (e.g. "do" or "ditto") is also not verbatim.
    – ACProctor
    Commented Dec 13, 2012 at 14:02

There are several aspects to any solution to this issue

  • a suitable value to go into the index;
  • how anyone would know the correct suitable value, particularly where children are concerned;
  • programming Ancestry's index handling software to work with the information that (in this instance) Selina is married to John. Without the programming, the suitable "correction" is useless.

It is at least possible to argue that "wife" should always be interpreted as "wife of the person on the line above". That way you don't need a separate value to go into the index. Except... this assumes that entries are written in that order - I don't remember seeing any deviation from that in a UK census, but I just looked at my US 1940 images, and the very first page I looked at had "Wife" first, then "Head". Still possible to program, but starting to make me worried about the possibilities.

I really like this issue and the lessons it has for searching - but I fear we are in "least worst" territory.

  • For all those pedants - yeah, that's people like me - who contend that "least worst" is the same as "best", I respectfully disagree - there's a feeling to "least worst" that "best" doesn't begin to cover!
    – AdrianB38
    Commented Dec 12, 2012 at 23:24

The indexing of the census is correct -- the original document records Selina Brady as wife and the indexing should faithfully reflect the underlying source, so I don't agree you should be able to amend it. However, I see you've suggested an annotation that clarifies the relationship, which is helpful to anyone viewing the same census page, and that seems the right way to proceed.

  • It only helps if someone investigates that record. All of my previous searches had missed it because the so-called accurate transcription of the data was extremely misleading. The scheme of adding alternative names & dates using - which do show up in searches - was obviously too difficult for Ancestry to make for relationships
    – ACProctor
    Commented Dec 13, 2012 at 18:10
  • The transcription was accurate. We're down to disagreeing about the interpretation of the evidence. I would rather work from accurate transcriptions than other people's interpretations. A search for Selina the wife of a Brad*y in Cheshire on Ancestry returns it in the top 4 results, which doesn't seem unreasonable. (It's 3rd if you don't specify Cheshire).
    – user104
    Commented Dec 13, 2012 at 18:27

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.