More about duplicate ghosts in the FamilySearch Family Tree

I've been waiting for some time to discover another set of "duplicate ghosts" in the FamilySearch.org Family Tree. Some of us have been finding these long lists of duplicates regularly. At the same time, representatives of FamilySearch are making comments about how the duplicate issue has been "resolved." The screenshot above shows an individual married to one of my distant cousins. There was the original entry from the Family Tree.

This project started with a routine correction of the entry for Mary Tanner by standardizing the entries of dates and places. That action started a cascade of changes resulting in obvious duplicates. Here is a screenshot of the 32 steps I have taken up to this point.

One of the interesting results was this second entry showing a number of children.

Bear in mind, that when I had reached this point the entry showed only one duplicate for John Briggs. I had already done only one merge. Before doing any more work, I decided to check for another duplicate. The screenshot at the beginning of this post was the result of that search. I did one more merge and when I returned to do a search for more duplicates I got the following list:

 The number of duplicates had increased from 5 to 8. This is what I mean by cascading duplicates. To trigger this kind of response in the Family Tree all you have to do is add a source and look for duplicates or make any other change and start looking for duplicates. Each time that you merge an individual, more duplicates appear. I have had this situation continue while I worked for four hours straight merging duplicates. At this point, a search for duplicates for Mary Tanner shows the following:

In short, there are currently 12 duplicate entries that need to be merged. However, as I proceed to do the mergers it is inevitable that more duplicates will appear especially for the children in this family. Past experience indicates that I may end up with dozens of mergers for this one family. None of these entries of duplicate individuals is shown initially when the user searches for duplicates. These are not hypothetical duplicates, they are real. Here's a screenshot showing the first merger screen:

I recently mentioned a post that I found 106 potential duplicates of one of my direct line ancestors. From my perspective, the duplicate issue is far from resolved. In fact, when I did the first merge listed above, I got the following additional duplicates for each of the children. This raised the number of duplicates needing to be merged to 20 after I had already done three merges.

In my experience, each succeeding merge will create additional merges for some time until this particular block of duplicates is exhausted or I am exhausted whichever comes first. The mere existence of these cascading merges indicates that there is a substantial reservoir of duplicate entries still waiting in the Family Tree.

I decided to do all of the merges before finishing in publishing this article. I kept track of the number of merges necessary, but after about two hours of merging records, correcting entries, standardizing dates and places, and adding in Record Hints, I did approximately 54 merges.

 When you encounter such a situation, you need to be careful to reload the pages constantly and recheck every time for more duplicates. In many cases, the duplicates are not found by the FamilySearch Possible Duplicates. You need to search by ID number and the duplicates are obvious. I am certain that it would take someone with less experience doing merges much longer to get through this morass.  I'm also certain that FamilySearch is vastly underestimating the number of duplicate entries left in the Family Tree. Here is what the family look like after all of the merges:

There were record sources for each of the children. There really were two daughters named Mary Briggs. After writing the above about the number of merges, I added a surname to the first entry for a child named Hannah and found one more duplicate.  The total number of duplicates turned out to be 55.