Try searching for MWIK ITIJGWAY for 1874.
There are 9 names listed but only 3 are real names.
Although these errors can be funny what we must realise is that we are paying for a service that we are asked to correct, how many other products do you repair once you have bought them ?

Thanks for your response. I can’t comment about Electoral Rolls because, as I mentioned, I can’t access them.

But I thought that my comments about Ancestry’s response to errors in census transcriptions might be relevent to the Electoral Rolls as well.

What blew me away was the fact that Ancestry expected me to correct the error myself by contacting the “creator” of the census data.

As for Member Solutions being unhelpful unless, as you said, “you can give them chapter and verse on a specific problem so they can verify it”, I don’t know how more specific I could have been when I contacted Ancestry. That is, in addition to the specific census, I also gave full details about the person omitted from the transcription (full name, age, est DOB, birth city and county, name of her school with specific town and county, district, census page number and page line).

Also, my apologies to everyone for adding comments which are not specifically about Electoral Rolls. It is frustrating that comments on Ancestry blogs are cut off after only two weeks of discussion rather than remaining open for anyone to add further comments as newly-discovered issues and problems arise, eg. recent 1911 census release as well as Electoral Rolls. Or, conversely, to give praise for a “job well done” when things are going well, errors have been corrected or a situation has otherwise improved.

But to have discussion closed off after such a short time frame, even when there are obvious and serious difficulties, it’s “Out of sight, out of mind. No problems here. Everyone move on.”

It is clear that a number of people have not been transcribed as they are on the same page and it is beyond belief that you ancestry have not listed whole boroughs as this is how the records are arranged by London Metropolitan Achives (LMA). I have confidence that LMA have supplied the whole collection, if not ancestry should have told us that!. Once again it appears that ancestry have transcribed records solely for financial reasons and not to help researchers!.

There is a similar missing persons problem with this collection but it is a little more confused.

I understand that the Electoral Rolls came from the London Metropolitan Archives as Microfilm and have been put through an OCR process to transcribe them by Ancestry. Ancestry have also supplied on screen images of the microfilmed pages.

When viewing a page, occasionally the OCR process appears to have skipped a person as they are not in the Ancestry Indax at the bottom of the view. I guess this problem is down to Ancestry to correct.

Every Elector on a Roll has a number allocated to them and in some cases it looks like whole page of the Roll are missing as numbers start at some odd number instead of 1. Now this could be down to Ancestry just not scanning the missing page or it could be that the LMA have not supplied it or it was never microfilmed in the first place.

I have found that Member Solutions tend to be a little difficult to deal with as unless you can give them chapter and verse on a specific problem so that they can verify it, they brush a complaint off. I have learnt that you need to persist.

There are also people missing from the transcriptions in various censuses. I’ve found almost a dozen myself, just stumbling across them in the course of my own research.
When I contacted Ancestry in October 2009 about one of these omissions in the 1871 census, the reply I got from someone at “Member Solutions” stated the following (and I think it may apply to *any* transcribed info on Ancestry, not just the censuses, which I why I’m mentioning it here) –

“…..Please note that we generally receive data from outside sources and post it as it was provided to us. In order to maintain authenticity in our records, we cannot make changes directly to material not created by Ancestry. To make these types of changes, you will need to contact the original creators of this data, have them correct this in their own records and then send this update to us for posting. We apologize for any frustration or inconvenience this may cause…..”

Note that this says that *I* had to contact the “original creators of this data”, not that Ancestry would do it. So, since I did not follow up on the above advice, the error/omission I told Ancestry about is still there in the 1871 census, ie. a researcher will not be able to find this “missing person” unless they know exactly where she was in 1871 (a pupil at a school in a county which was not her home county).

The reply from Ancestry also stated –

“….If you wish to consult our corporate policy regarding the content provided on Ancestry.com, please go to the following URL:

http://www.ancestry.com/legal/terms.htm#Liability

…..”

It also went on to state –

“…..For additional information on the source of the database, select the Database Title link found at the top of each page of search results. The next page provides information on who compiled the database, including a short bibliography and a brief description of the information contained in the database. Here you will be able to locate the source information for the database in question. This should assist you in contacting the source directly on this matter.”

Perhaps someone can see if this applies to the Electoral Rolls as well (which I can’t access because I don’t pay Ancestry enough for my subscription).

Lastly, Ancestry finished off with this very unhelpful bit of advice –

“….You can also typically add a User Comment to any record that you feel may be incorrect. You will find a “add alternate information” or “add comment” link to the right of each database result. If you add a User Comment (once the feature is again available), your comments or corrections will be connected to the database result for others to view.”

But I had not contacted them about something which was *incorrect* — it was about info that was missing altogether. The above strategy will only work if there is an individual listed in the transcription to add a correction to — it’s useless if the person isn’t listed in the transcription at all!

Good luck correcting names, there are over 139 Million of them to go through in this collection!

The silliest thing is where the OCR scanner has been run over a header page and has attempted to create names and address from it. Follow this link

http://search.ancestry.co.uk/Browse/view.aspx?dbid=1795&path=Haringey.Hornsey.1937.432&sid=&gskw=

and laugh at the 10 voters listed in the Ancestry Index at the bottom of the screen. This is just one of many examples.

How can Ancestry release this rubbish!

By browsing the Collection at

http://search.ancestry.co.uk/search/db.aspx?dbid=1795

you can get a dropdown of the Boroughs/Counties, Parliamentary Divisions and Years that have been transcribed.

A quick look shows, for example, numerous records allocated to the borough of Barnet. Everyone that I have so far looked at should have been in Middlesex. Strangly Middlesex appears to have only one year allocated to it.

Random sampling of other Boroughs seems to show that just about every one is wrong! Ancestry is not even consistant with the mistakes allocating records to a name Borough for some years or simply London for other years.

What a mess – and its all caused by using OCR and not checking the output. Or maybe just having someone who has no knowledge of London do the checking.

The interesting thing is that every record I have looked at clearly includes all the information necessary to get the index right in the first few images.

What

The basic question is why a list of the boroughs was not prepared before starting and what parishes were included and when, London has had several borough changes over this period. Surely making wide-ranging comments about 130 years that we should be told what rolls have been completed and over what period. I am fed up with receiving ancestry customer satisfaction survey e-mails because the answer is the same in that ancestry have not dealt with the issues.

Basically, Ancestry used OCR technology to transcribe these records and dispite the numerous obvious errors seem to be proud of the result. What appears to be a standard e-mail I received states -

“Because these Historical Records have far too much text to be indexed manually, indexes are built using advanced Optical Character Recognition (OCR) technology to automatically recognize text within an image. We believe we are using the very best OCR technology available, but it is still less accurate than human eyes and brains. These limitations make it impossible to achieve the same level of accuracy found in other indexes. In cases where the original microfilm text is unclear, the OCR technology tries to guess words intelligently.

There are two important points about searching the Historical Record images:

1. It is likely that there are additional matches for your search that could be found by browsing but which are not in the index or search results list.

2. The search results may include “matches” on words other than those you searched for because the Optical Character Recognition (OCR) technology we use believes they are optically similar enough that your target word(s) may be in the original image.

I do know that the content team is aware of this issue and will be fixing the records as soon as possible.”

As for the problem with Ancestry using Boroughs that did not exist at the time of the record, well this was ignored.

One other comment in the e-mail indicated that Ancestry hope to double the number of records available. No firm date was given but “Fall” 2012 was mentioned. This tells me two things, we are dealing with Americans which perhaps explains the problem with Boroughs and that this collection is incomplete. I have asked what percentage of the total records is currently available but I have yet to have an answer to this question.

Overall I still feel Ancestry could do a lot better.