Saturday, May 18, 2024

When the Calculations are Slightly Off


I have an ongoing tiff with the ThruLines gang at Ancestry. Yes, I'm quite sure that "gang" is really a collection of computers hazarding guesses about family relationships based on majority rule—the more trees that say so, the truer it must be, right?—but let's call it my little lovers' quarrel, nonetheless. The point being: no matter how much I appreciate the ThruLines tool, those calculations are slightly off. And it bugs me.

I decided to get over my petulant ways and go back to review the ThruLines matches for my mother-in-law's fourth great-grandparents, William and Elizabeth Ijams. Last time I visited that neck of the Ancestry woods, they had assured me that my husband had fifty six DNA matches linked to William Ijams as a common ancestor, and fifty eight attributed to William's wife Elizabeth. Only problem was: when I open up the report for either William or Elizabeth, there are matches attributed to "children" of the couple whom I somehow haven't been able to identify as their children, despite serious searching through documentation.

In other words, I didn't think those names belonged in that branch of our family tree.

After reconsideration, I decided to take a closer look. After all, I have missed entire branches of that tree before—and mended my ways specifically because of what the DNA test told me.

What I saw, when I examined those possible names a bit more closely, was that the lines of descent offered up by ThruLines calculations sometimes contained names that I did have in my mother-in-law's tree. Only problem was that I had those names in different places, For instance, one DNA match had Richard Daniel Ijams listed as son of William and Elizabeth, when he really was son of a different William: the original couple's own son by the same name.

I can see how someone could make such a mistake. Seeing the same name in two separate generations can happen in a lot of cases—a father naming his son after himself—but there are other contributing factors that may come into play, as well. In this case, the senior William had specifically mentioned Richard in his will, but he also noted that Richard was his grandson, son of his own son, William. Perhaps because the will was transcribed in such a way as to separate the word "grand" from the word "son," someone may have missed that fact and entered it into an online tree incorrectly—but was that mistake replicated so many times as to fool the algorithm into thinking it was correct to skip an entire generation?

I decided to go back to all those DNA matches which ThruLines had misattributed to the wrong ancestor, and rework the suggested tree from the most recent descendant I could find in my tree. After all, I've spent years building out the collateral lines in this family; at some point all that work is going to pay off. 

In this case, it is. I'm updating and making the connections—hopefully outlining the right path to the most recent common ancestor for the record. Perhaps at some point, the genealogy tide will turn and those busy tree-copiers will pick up my lead and replicate the right Richard Daniel Ijams in their own trees. Give us time—and some convincing documentation—and we'll outnumber 'em.

No comments:

Post a Comment

Related Posts Plugin for WordPress, Blogger...