Linking Genealogical Evidence: A Method – The Global Tofay

Being able to accurately and reliably link evidence from different sources to the same individual is a key skill which all family historians need to learn and master. But there is very little writing out there to help people learn how to do this – and the vast number of poorly reasoned linkages in member trees at Ancestry and elsewhere is testament to the need for broader discussion on this topic.

Linking evidence from two different sources to the same person is one of those tasks in genealogy that is fundamental, it underpins the collection and collation of all evidence and is essential for turning a body of evidence into a genealogical proof. It is so fundamental that we often do it without devoting a great deal of thought to how we do it. When is it okay to accept that two records relate to the same person and when is it not?

Factors Underpinning Evidence Linkage

In the early years of my family history research I had developed a kind of subconscious framework when deciding whether two pieces of evidence were referencing the same person. Much of my research was centred on small, rural communities in the north of England, so it was not difficult to decide that Daniel Usherwood, a husbandman, who married in 1750 was the same Daniel Usherwood, a yeoman, who married a second time in an adjacent parish four years later, his first wife having succumbed to the perils of 18th century childbirth. However, it was much more difficult to tease out the records relating to another relative, James Green, who was baptised, married and buried a little later that century in the nearby market town of Warrington, Lancashire. There seemed to be lots of men by the name of James Green, all of whom were labourers. It was difficult, verging on the impossible, to distinguish between them, so doubly difficult to link evidence together into a meaningful story for my research target.

So why was one case so much easier than the other?

I think there are five principal factors that define the ease or difficulty of linking two pieces of evidence together:

1. Uniqueness
2. Community Size
3. Distance
4. Time Difference
5. Contradictory Evidence

Let me explain in a little more detail:

1. Uniqueness

We all know that it’s easier to research an uncommon name than a common one. That is because the more uncommon the name the greater its uniqueness. In my example above, in the mid-18th century Daniel Usherwood was an uncommon name, but James Green was not.

Uniqueness is defined by more than just an uncommon name though. When comparing two records we should extract all the common information that appears in both records:

Marriage 1:

Marriage 2:

The common elements are: Name, Abode, Approx. Age, Occupation (husbandman and yeoman are both small-scale farmers), Literacy.

The combination of name, abode, age, occupation and literacy represent a strong, unique “signature” for this person. On a rating scale of Low, Medium and High uniqueness, we would have to rate this as High.

2. Community Size

As all family historians know, it is easier to trace the life of a person who lived in a small community, like a rural parish, than someone who lived in a large community, like a city. The possibilities to confuse two different people multiply rapidly in larger communities. With larger numbers of people there are much greater chances of two people sharing similar names, born at similar times, who held similar occupations.

In the case of Daniel Usherwood, resident of the parish of Frodsham, Cheshire, the community size is Low.

Other relatives of mine living in the nearby market town of Warrington, Lancashire, were in a Medium sized community. Yet others, in the nearby, booming city of Liverpool, were in a High sized community.

3. Distance

The geographic distance between the two records you are trying to link is also a factor. Our confidence to relate two records in the same rural parish, or perhaps an adjacent parish, will naturally be higher than when the recorded events were separated by tens or hundreds of miles.

In the case of Daniel Usherwood, who married first in Frodsham, then in the adjacent parish of Runcorn, the distance between the events is Low.

In my case study The Woman Who Fell From The Skies (see I traced the movement of Mary Jane Hyland who married and lived her adult life in Warrington, Lancashire, but had been born 20 miles away in Liverpool. That distance would count as Medium. Her parents, William and Margaret Hyland, had migrated from Ireland to Liverpool, so the distance between records for them would be High.

4. Time Difference

The time difference between two records we are trying to link is another key factor. We will naturally find it easier to link together two records that take place a small number of years apart than records that are, say, 30, 50 or 70 years apart.

In the case of Daniel Usherwood, who married first in 1750 and secondly in 1754, the time difference is Low.

5. Contradictory Evidence

No matter the uniqueness of the data, the community size, the geographic distance or the time difference, if the data between the records is sufficiently contradictory then we are likely to conclude that they do not relate to the same person.

In Daniel Usherwood’s case, his occupation in 1750 is given as a Husbandman, while in 1754 he glories in the (slightly) grander title of Yeoman. They are distinct but related terms for small-scale farmers – a yeoman is generally considered higher status than a husbandman as he owns his own land. These aren’t incompatible terms for Daniel, he may have acquired land since his first marriage in 1750 (perhaps in his marriage settlement) or may simply have given himself a grander title out of vanity. For Daniel the contradictory evidence of subtly different occupation titles across the two marriages would rate as Low.

If one record said he was a blacksmith and the other record that he was a carpenter, then the contradictory evidence would have been High – these are not interchangeable terms and are both trades that required training through an apprenticeship.

However, I have a concern about two marriages in swift succession. On their own, these two records provide only slim evidence of Daniel’s first wife having died in such a short time window, between June 1750 and July 1754. The second record says that Daniel is a widower, which is perhaps indirect evidence for the first wife’s death, but I feel the need for something more concrete. As such, I will rate contradictory evidence as Medium.

Write Down Your Assessments of the Five Factors

After assessing the two pieces of evidence according to the five factors above, list your results in a small table:

Next Time: We’ll look at how to assess linkage strength and subject it to rigorous tests.