Monday, March 14, 2016

Monday Mailbox: B. Canada is Not British Columbia

The Ancestry Insider's Monday MailboxDear Ancestry Insider,

Robert Henry Kittson was born in what is now the province of Quebec, Canada.

The Ancestry.com indexer of the 1881 census of Sorel, Richelieu County, Quebec, misinterpreted his birthplace, recorded in French as "B. Canada," (abbreviation for "Bas Canada" i.e. Lower Canada) as British Columbia. I was surprised by this, checked the image, then the onscreen index, and lo and behold, everyone recorded as born in Bas Canada, had been interpreted by the indexer as having been born in British Columbia.

Ancestry needs to review the instructions it gives to indexers.

I have corrected the error for all the names on that image. A search for anyone born in British Columbia and living in Sorel in 1881 comes to 1,593 people! This kind of error is egregious, and correcting it should NOT be the job of Ancestry's paying customers. Ancestry needs to make a global correction of this error.

It is also apparent that whoever was doing the indexing had NO knowledge or background regarding Canada's history or geography. Surely it should have occurred to him or her that it was extraordinary that hundreds of people had been born in British Columbia yet by 1881 they had moved to Sorel, in what is now Quebec?

Signed,
Jean F. Milne

Dear Jean,

Since Ancestry.com employs cheap, offshore and sometimes non-English speaking, labor to key its records I believe it employs the rule of “key what you see” with its indexers, precisely to avoid this kind of issue. Database programmers then detect names that don’t match standardized places. The database programmer writes a mass update instruction that replaces all occurrences.

A single database instruction created this mess and a single database instruction can fix it. Hopefully, the people at Ancestry who read this newsletter will inform the appropriate people to have that done.

But why did a database programmer create it in the first place? Genealogical experts warn us that to properly interpret a record, we must first understand it. For example, Elizabeth Shown Mills has written,

Accurate evaluations of evidence require researchers to have a sound technical knowledge of the materials they use. We cannot expect to pull a census and scan names or run statistics without thoroughly understanding the circumstances under which that record was created. [Several factors, including] the abbreviations that differ from modern usage [or in this case, the reader’s own language]…all affect our interpretations.1

Ancestry and FamilySearch have excellent, programmers and publication personnel with lots of genealogical experience. But they need to approach record experts whenever there is any doubt, whatsoever, that they can properly interpret a record.

Elizabeth’s warning applies to each of us as well. I’m a big believer in education. There is always something more to learn that will help us as genealogists. (The more I learn, the more I realize how poorly I measure up to the knowledgeable genealogist I have the honor to work with in this community!)

Signed,

The Ancestry Insider


Sources

     1. Elizabeth Shown Mills, Evidence Explained: Citing History Sources from Artifacts to Cyberspace, third edition, Adobe Digital Edition, (Baltimore, Maryland: Genealogical Publishing, 2015), 21.

16 comments:

  1. Another problem this highlights is the mass exclusion of librarians for computer scientists. This is seen in many areas of genealogical effort.

    ReplyDelete
  2. The indexers are probably paid by how fast and how many records they finish. Hence, stupid mistakes are made of the most obvious spellings, names, locations and facts. If you are going to hire people to do a job at least make sure they do it correctly,since the index is the most common way we have to locate a record.

    ReplyDelete
  3. The indexers are probably paid by how fast and how many records they finish. Hence, stupid mistakes are made of the most obvious spellings, names, locations and facts. If you are going to hire people to do a job at least make sure they do it correctly,since the index is the most common way we have to locate a record.

    ReplyDelete
  4. "Key what you see"--if you have no knowledge of English, you can't key what you see, because an unknown language makes no sense to your eyes. I have seen transcriptions of directories and voter's lists that were so infuriatingly wrong, in every line, for name, place and occupation, that it would be impossible ever to correct them--the names and places being pretty much what you might expect if you picked up letters at random out of a Scrabble box. You could not possibly correct these sorts of errors with a key stroke. I have reported the issue to Ancestry many times--as a correction on the relevant page--and have never heard one comment or explanation from anyone there. I think they are happy to have us do the real work of correcting the pages for free. However, that assumes you can find it in the first place. Which, with the botch made of many of the names, would quite often be simply impossible. I might also point out that Ancestry has yet to correct--with a keystroke or otherwise--"Canada, England" as a POB on many Canadian-born people on US census records. This is a common misreading of "Canada, Eng"--US census-speak for "English-speaking Canada", and a ridiculously common misreading. There is no place in England called Canada, but the mistakes keep right on coming.

    One gets so tired of doing Ancestry's work for them, and with so little evidence that they intend to change anything, appreciate your input, or even will acknowledge the issue. But hiring foreign transcribers is the very worst thing they have ever done. A travesty.

    ReplyDelete
    Replies
    1. Judy, the city directories and voters' registration lists are not keyed by humans, but by OCR scanning. That's why they're so scrabbled.

      Delete
    2. Thanks for the information. but if they are so truly awful, which they are, why would they persist in foisting incoherent rubbish on us paying customers?

      Delete
  5. Check out the 4,000+ people living in Philadelphia in the 1870 census who show born in the Phillipines because their birth place was shown as "Phila."

    ReplyDelete
  6. Oh my gosh--that is (almost) hilarious!

    Why does there seem to be utterly no oversight or mechanism for spotting even egregious errors? Ancestry Insider, do you know how it is possible that utter nonsense can be published?

    ReplyDelete
  7. How about all the people born in the state of Georgia (USA) listed as being born in the Nation of Georgia? Garbage in garbage out.

    ReplyDelete
  8. This example of transcription error in Ancestry.com records is one of too many. I and others spend more time than I'd like, trying to point out errors in Ancestry records. Too often, looking at the image shows something different than what Ancestry says is in the record. Marriage intentions become "marriage date;" a death becomes "residence;" a listing of a widow's late husband in a city directory becomes a "residence" for the husband; a burial date becomes a death date, etc. Whenever there is no image, I am leery of accepting Ancestry's account of the record. It seems to me that the problem has gotten much worse lately.

    ReplyDelete
    Replies
    1. I haven't used a transcript from ancestry for more than a couple of years. And NEVER an "index only". I no longer pay them for anything. I can use their hints and find the answer somewhere else.

      Delete
  9. When researching my mother's maiden name "Trucks," I always check indexes under "F." Keying what you see, I guess...

    ReplyDelete
  10. The worst one I've found lately is that all the Drouin records for St-Felix-de-Valois are really records for St-Felix-de-Kingsey! I can live with errors in the indexing, but when they totally misidentify the source we are supposedly looking at???

    ReplyDelete
    Replies
    1. Well, turns out I'm wrong! The church in Kingsey is called St-Felix-de-Valois, although FamilySearch lists it as St-Felix-de-Kingsey, probably to distinguish it from another St-Felix-de-Valois, which is in St-Felix-de-Valois.

      Delete
  11. I don't have an Ancestry subscription but have they moved Prince Edward County out of Prince Edward Island yet and back to Ontario where it belongs? And how many people with the surname Ditto are still listed?

    ReplyDelete
  12. Another Ancestry.com transcribing error I discovered is when they index E. T. City in Tooele County, Utah as "East Tooele City" for the 1860 US Census when it should just read as "E. T. City" because E. T. stands for Ezra Taft, the given name of early Mormon leader/pioneer Ezra Taft Benson (1811-1869). E. T. City is now known as Lake Point, Utah. I believe this name change occurred just before 1900.

    ReplyDelete