Research Help section can be powerful tool, or lead to records corruption
I have been part of Family Tree BETA groups which test new features for years - for a very long time I wished it was easier to attach sources (particularly census records) to multiple individuals in the family and provided feedback to that end. The new Research Help section enables quicker documentation of vital information for families. Unfortunately, it's also leading to widespread corruption in family records. Patrons who are not well trained in using Family Tree (or just generally careless) are using the "create new person" feature to add children to families for which there is no relationship. Example #1: One couple, Michael and Bridget Donnelly, have two children - John and Mary, found in the 1900 census. Another couple, Michael and Bridget Donnelly, have six children - Michael Jr., John, Mary, Patrick, Anna, and James, found in the 1910 census. Further research demonstrates they are not the same families. Patrons assume, however, they are the same and add each additional child from the second family to the first using documentation found in Research Helps. This leads to erroneous merges and additional mistaken sourcing down the road, as well as confusion regarding mother Bridget's maiden name. Example #2: Michael and Bridget Donnelly are confused with Michael and Bridget Donnell (no "y" at the end). Since most of the said patriarchs and matriarchs of the above examples originated in Ireland and migrated to the northeast United States, extensive research must be done on each and every family to ensure records and relationship integrity. I have spent the last few months sorting out families that were corrupted using Research Help. I'm up to 53 Donnelly/Donnell/Donerly families and 11 Connelly/Conerly/Conerty families that are now repaired, but it's beyond disheartening. In an ideal world, patrons would simply proceed with more caution. Realistically, I believe stricter algorithms on the backend are in order. The problem isn't nearly as rampant with Ancestry.com because more stringent examination of the actual records is required.
Comments
-
I think the problem is just as rampant on Ancestry. It may be less obvious because it's all individual trees with conflated Donnelly families instead of a single tree with conflated Donnelly familes. As a Donnelly researcher, on FamilySearch, Ancestry, and many other platforms, I have much experience on the subject.
1 -
It is just as rampant on Ancestry, @Áine Ní Donnghaile . When you reject a hint, you can indicate what was wrong: names, dates, etc. I have done that sometimes but I'm not sure that information is wrapped back into the logic. When I have 2 or 3 generations of men with the same name, nothing seems to stop all hints appearing in all generations, even when ridiculous. Grandpa died an old man in 1858, dad died an old man in 1880, and Jr served in the Civil War and collected a pension until dying an old man in 1920. The pension record will be hinted in all generations.
1 -
There are 4 new hints for my GG Aunt (not a Donnelly) on Ancestry. Sarah Anne McGarrity was born in 1870, in Ireland, and died in 1888, in Philadelphia. She's been documented in my tree with her birth and death records for at least 12 years.
Of the 4 new hints, not one is for a woman named Sarah Anne. One is a passenger manifest showing "Hannah McGority" arriving IN ENGLAND in 1898. One is for "Anne McGarrity" aged 20 in 1870. I guess I should be glad it's at least in Philadelphia, but Sarah Anne wasn't in Philadelphia yet, and she certainly was not aged 20.
0
