1911 CENSUS (General)

by Jefff , West London, Middlesex, Sunday, May 20, 2012, 14:54 (4573 days ago) @ mcowan

Thanks Anne for your helpful reply, particularly your views re the various subscription sites. The only reason I've leaned towards Ancestry is the free access available at the local Libraries, albeit limited, plus as you say they do seem to have the greatest records. You are quite correct wrt the FOD etc transcribers using their local knowledge to help ensure accuracy, which is great, but what does get my goat is the ability of Ancestry in my limited experience to get the simplest record so wrong. The number of Marry's instead of Harry's I've found are an obvious example. Last week I was excited to find an ancestor who had been a "ship steward", great I thought (I'm an engineer). However the 1911 Census form in question showed this person to be living with a Farmer, in rural Longhope. Yes the Form clearly read "Shepherd" !!.

Further to my post of last week I have today found the following article from 2006 within the website of a wellknown British PRs company, which gives some credence to my thoughts:

"??? Genealogy Supplies’ online genealogy data service has taken off, and they are striving to make their indexes as accurate as possible, largely helped by the UKIndexer project www.ukindexer.co.uk. This is part of a vigorous, many layered, checking system to ensure a high degree of accuracy.

Transcripts and indexes
Although they started out with a volunteer surname indexing project, the census pieces are now fully transcribed in India with good results. The transcription stage is followed by an initial check to sample the data, which is then rejected if the error rate is unacceptable.

The data is then examined by a series of sophisticated, custom written software programs which look for anomalies. Initially these look at names and their frequency, flagging unusual looking entries. The flagged entries are checked and corrected by the ??? Genealogy Supplies trained in-house indexing team and the end result is used to create the name indexes. These are released in advance of the full transcripts, which still have a considerable amount of work to be done.

Further customised software examines the fully transcribed data, which contains all the major fields including the occupation. This software looks at details such as place names, relationships and occupations, again flagging any entries that look dubious. These are then checked by their in-house team assisted by over 4,000 volunteer indexers who have signed up to help at www.ukindexer.co.uk."
http://www.genealogyreviews.co.uk/ftmMar06uki.htm

Please note I am not promoting this company in any way, however it strikes me that their multilayered process is rather more effective than the system Ancestry use, which I'm sure is similar wrt their using staff from countries who do not fully recognise or understand British names and their context.

With regard my previously posted guess that the transcriptions would be initially done by machine, I mentioned this issue to my longsuffering wife. She is an accountant, she tells me in the last few weeks she's had to contact banks who are incorrectly removing extra £000s from her company's account. It transpires my wife has a new clerk who writes cheques very neatly & clearly; however her "1"s have long leading strokes which make the automated readers think they're "4"s !, rather like the Victorian "H"s and "W"s which Ancestry often think are "M"s.