Wonderful Albert-Casey-itis

After my last post about mistrusting databases, Dermot Balson commented with a discreet little link to a chart of his that deserves to be brought out and displayed proudly to the wider world. Here it is:

It charts the number of people recorded at each age in the 1901 and 1911 censuses and confirms beautifully what most researchers in the censuses have long suspected. The rounded years (30, 40, 50, 60 and so on) are wildly over-represented – most people born before 1870 simply didn’t know their age. Not surprising, given that they never saw a calendar or celebrated a birthday.

Look at the chart longer and other things become clear. The drop in total population between 1901 and 1911.  The bumps around the mid-decade ages (35, 45, 55 …), indicating that many of these were also just guesstimates. And my favourite, that improbable leap in numbers aged between 70 and 80 in 1911. The Old Age Pension was introduced in 1908 for people over 70, immediately making it very important to be at least that age. By the look of it, at least 30,000 individuals promptly suffered accelerated aging.

I contacted Dermot to ask his permission to use the chart and also if he had more. Look at a sample of what he came back with:

Causes of death, Mourne, 1864-1921
Age distribution of deaths, 1864-1921
Literacy 1861-1921
Pregnancy timing by occupation 1864-1921 (a proxy for seeing when different occupations were busiest)


The latest YouTube video talks through all of these.

Dermot suffers from an advanced case of Albert-Casey-itis, where a researcher (usually descended from an Irish emigrant) runs out of ancestors but can’t stop.  So he moves on to his ancestors’ neighbours, then to his ancestors’ neighbours’ neighbours, then to the entire locality … A truism of Irish research is that the border between genealogy and local history is very flimsy.

Like Casey, Dermot concentrates his record-collecting on the area his ancestors came from, the Kilkeel area in South Down. But as you can see, what he does is vastly superior to the Casey  pile-em-high approach. A retired actuary, his spreadsheet skills are awesome. As he explained to me, he collects “transcriptions of all record sources in a single spreadsheet (with over 40 sheets), standardizing names so you can actually find records when you search, and linking births, deaths, marriages and censuses together, so that given a name, I can immediately find their marriage, a list of their children, their census records, any family deaths, newspaper references, and parent information where available, each of those with direct links to online scans.”

The end result will be the ultimate local and family history resource for the area. As a sample, Dermot has sent me a screenshot of part of his Excel file:

Master spreadsheet

He’s anxious to spread the gospel far and wide and to share his data- here’s an extraordinary downloadable PDF of his analysis of Mourne death records.

There is at least one doctorate here for someone collaborating with him. Any Irish third-level institutions interested?

Beware Mr Smarty-Pants Database

The most common mistake made when starting research online is surprisingly counter-intuitive: too much precision. The fact is, the more detail you include when you query a genealogical database, the less likely you are to find anything useful.

Just think. You know your Michael Barrett was born on March 17th 1868 (March 17th, St. Patrick’s Day, being the most common birthday for the nineteenth-century Irish), married Mary Murphy on September 14th 1890, had children John (1891), Mary (1893), Michael (1896) and Delia (1898), and lived all his life in the townland of Ballybeg, Co. Mayo. Carefully enter all of this  into a search form and you are guaranteed to find nothing.

All it takes is a single non-matching item: in the originals, (the page recording Michael’s birth was used to light a fire in 1898); in the database transcripts (the transcriber had a late night and dozed off over the marriage record); a single item misreported by the family (John was actually born in 1890 and there are four  Ballybegs in Mayo). The response from Mr Smarty-Pants Database will be the same for all: No Match.

Nothing like some birth records to get a good fire going

Don’t get me wrong. Knowing these details may eventually provide evidence to unlock the truth. But to start off, you need to cast the net as wide as possible. How many Barrett births are registered in and around all of those Ballybegs between, say, 1864 and 1870? How many Michaels? Can you identify the precise marriage registration, using only the names, not the reported date? What are the ages given in the 1901 and 1911 censuses? Do they match each other, or the ages you think you know? (Unlikely.) Are there other Barrett households in and around Ballybeg in 1901 and 1911? Any with heads of household of an age to be siblings of Michael ?

The biggest sites – ancestry.co.uk, FindMyPast.ie, irishgenealogy.ie – all funnel their users like this, starting off broad and ending narrow, because it’s by far the most productive way to use their records. Their search interfaces force you into it. They know what they’re doing. They know their databases are stupid.

If you’d like an even more ranting version of this, I’ve gone on (and on) over at YouTube.