Friday, August 18, 2017

Skepticism and the Census

Anyone who has spent much time chasing their ancestors in the census knows that you have to take the "facts" on a census return with a few fistfuls of salt. A lot of our forebears, especially the ladies, seemed to view "age" as more of an abstract concept than an objective fact, either not knowing, not caring, or actively lying about their age (how many of yours, remarkably, aged less than 10 years between consecutive censuses?).1

Then there's the reality that, except for the 1940 US census, we have no idea who actually answered the enumerator's questions – perhaps a step-parent, a 12-year-old, or a handy neighbor. Good luck with getting valid information there.

And of course, there are the enumerators themselves: they sometimes missed individuals2 or whole families,3 applied great creativity to the spelling of names,4 often had atrocious handwriting, and no doubt compounded these problems with copying errors.

But did you ever consider whether the pages were even numbered in the right order? I'm not talking about pages being microfilmed or digitized out of order, or missing pages – that's a whole separate issue. And I don't mean the often multiple numbers confusingly stamped or scrawled on pages after whole cities or counties were assembled. No, I mean when the pages are actually numbered 1, 2, 3, and so on, within a single enumeration sub-district or similar entity, and appear in that order on the digitized microfilm. I always pretty much assumed that could be relied on.

Until the other day, while I was working on my Saint-Cyr line in Quebec's Eastern Townships.

For those who aren't familiar with the 1861 census of Canada East (Quebec), it's a fully nominal census (i.e., all household members are named, not just heads of household), enumerated on 50-line sheets. However, there are no dwelling or family numbers, and no relationships to the (unidentified) heads of households. You have to deduce where a dwelling begins from the columns describing what the house is built from and how many stories it has, and where a family/household begins by how many families live in the house and who has an occupation listed. To further complicate matters, the enumerators were inconsistent (at least where my folks lived) about the order in which they listed family members after the apparent head and apparent spouse: sometimes they're in decreasing age order; sometimes boys first, followed by the girls, or vice versa; if a household member was absent at the time of the census (there are columns to indicate that), they might be listed at the end of the household, or mixed in with the others in whatever order they were in.

With that background out of the way, there were two "Sinsire" households (a father and an adult son) in district 2 of Arthabaska Township, Arthabaska County,5 the father's household (all surnamed Sinsire) occupying lines 45-50 at the bottom of page 3 and the son's (mixed Sinsire and Bélanger, from the wife's previous marriage), lines 44-50 at the bottom of page 7. Of course I checked page 4 to see if there were any more "Sinsires" at the top – there weren't – and page 7 was the last page of sub-district 2, so there were no more in the son's family. I thought.

Then I moved on to the 1871 census, where I discovered the son had two additional children, both well over 10, who should have been there in 1861, but apparently weren't.6 Was there a missing page 8 in Ancestry's database? (There was no page 8 or 9, with district 3 starting at page 10.) I checked on Library and Archives Canada. There was no sign of a page 8 there either. I checked that specific district for "Sinsire" entries, and, oddly, turned up two additional names, neither of which were the missing ones from the son's household. They were aged 14 and 11, listed on lines 1 and 2 at the top of page 6... but the last household on page 5 started with the surname Croteau on line 50. I began to smell a rat.

One of the missing children from the son's family was named Felix, so I searched for the first name in that district. Lo and behold, there was a Felix Bélanger – I had been looking for a Felix Sinsire, because in 1871 he was listed with the surname St Cyre – as was everyone in the household including the known Bélangers. Along with Felix Bélanger in 1861 was Edèse Bélanger. My other "missing" child found in 1871 was Thedèse St Cyre. And both of these kids were exactly the right ages. That was when I realized that they were too old to have been St Cyrs: they had been born before their mother had been widowed and remarried. Like the two mysterious extraneous "Sinsire" children, the two Bélangers were at the top of a page, in this case page 3, while page 2 ended with a Bilodeau family.

In short, the seven pages for the district had been numbered 1 to 7, but in apparently random order. By examining the names at the tops and bottoms of each page, I determined that the proper order of the pages is as follows:
  • 1 – Laliberte ... Barbier7
  • 7 – Turgeon ... Sinsire [fils]/Bélanger
  • 3 – Bélanger ... Sinsire [père]
  • 6 – Sinsire ... Frenette
  • 4 – Frenette ... Perrault
  • 2 – Perrault ... Bilodeau
  • 5 – Bilodeau ... Valliere/Croteau (2 families, 1 house)
Once I had the pages in the right order, all became clear. The two Bélanger children were the "St Cyre" children in the son's family in 1871, and the two extra "Sinsire" children belonged to the father's family.8

I have no idea if the rest of the districts – or the rest of county or province, for that matter – are as confused as this one small chunk. But I'm sure going to keep this in mind for the future: checking the "next page" for additional family members may give you a false negative – the page numbers may simply be wrong, and those additional members may be on an unexpected page.


SOURCES
  1. I have at least one who miraculously managed not to age at all between two censuses. 
  2. My father and his sister, aged 8 months and just over two years, respectively, were apparently not in residence with their parents when the census taker came around in 1920. Where else they would have been in January, in Maine, I can't imagine. I also can't imagine my grandmother just forgetting to mention them.
  3. Like, say, my mother's entire family in 1930.
  4. I can see maybe having trouble with an uncommon name like "Sukeforth," but you would think it would be hard to foul up "Kirk." You would be wrong.
  5. 1861 census of Canada East, Arthabaska County, Quebec, district 2, Arthabaska township, pages 1-7; database and digital images, Ancestry (https://www.ancestry.com : accessed 9 Aug 2017).
  6. I also had some issues with matching up daughters in the father's family, but three quarters of Catholic French-Canadian girls are named Marie something-or-other, and the names they actual go by are as malleable as their ages, so it can be pretty hard to tell who's who from one census to the next. Plus, they could have gotten married and left home.
  7. Or, this page could belong at the end. It's the only "standalone" page that can't fit between any two of the others.
  8. At least one of the names even matches up to a known St Cyr daughter (Loise = Marie Eloyse). The other one was listed only as Marie in the 1861 census, so she might correspond to just about any of the younger known daughters.