Genealogy from the perspective of a member of The Church of Jesus Christ of Latter-day Saints (Mormon, LDS)

Wednesday, August 3, 2016

Is Accuracy an Issue in the FamilySearch Family Tree?

In any engineering project, there is always an allowance for an acceptable or reasonable margin of error. From time to time, I have heard those working on the FamilySearch.org Family Tree refer to this margin of error. There are a lot of different ways that the Family Tree may be subject to error.

As users we focus on the actual entries in the Family Tree and evaluate them according to our personal understanding of the amount of "error" we find. However, our own standard for measuring the error of others is seldom a concern. I constantly find that people assume, rightly or wrongly, that anything they contribute to the Family Tree is "correct" and any variance from their contribution is automatically "wrong." Of course, some of the time, they may be right, especially if their contributions are bolstered by supporting source citations.

When the engineers speak about a Family Tree margin of error, they are talking about something entirely different. They are primarily concerned with programming issues that may or may not produce the results that are expected. For example, presently, FamilySearch claims a high degree of accuracy for their Research Hints, mostly I have heard claims upwards of 98% accuracy. However, this is just another way of stating the margin of error. Assuming you do have a rate of 98%, that means there is at least a 2% error rate. If this 98% rate is correct, then the program is really amazing. But how do you (or anyone for that matter) measure this accuracy? Let's assume that the claimed rate of accuracy is 98%, then we should be able to click on 100 Record Matches and find that they applied to the designated ancestor 98 times. However, that is not how the whole thing works.

Considering the fact that the users of the Family Tree play a major part in assuring the accuracy of the entries, most of the engineering issues of accuracy do directly affect the accuracy of the information in the Family Tree. Even if the search engine accuracy varies and a user gets slightly more or slightly less accurate Research Hints, ultimately, the user is the one who decides whether or not the hint applies or does not apply to the relative.

So, when the user considers the accuracy of the Family Tree, the main and most important criteria is the accuracy of the entries. Even if FamilySearch has made a search and given the user a Record Hint, the ultimate decision to add the information from the Record Hint lies with the user. Therefore FamilySearch in reality  bears no responsibility for the accuracy or inaccuracy of the entries. Any inaccuracies in the Family Tree were put there by the relatives of that particular person. Any changes to the Family Tree attributed to FamilySearch involve the work done by someone in the deceased person's family. Even in the case of the suggested entries that came from the International Genealogical Index or IGI, the data in the entries came from the user's relatives or from the original records.

When you are looking at the Family Tree and see something that you don't agree with or that you feel is wrongly recorded, you have just as much responsibility as anyone else in the world to correct the entry.

2 comments:

  1. Until the church starts emphasizing good research instead of "find a name", the familysearch tree will be full of errors.

    ReplyDelete
    Replies
    1. I don't think the accuracy of the Family Tree has anything to do with what FamilySearch says or doesn't say. The Family Tree is a wiki and will, by its nature, become more accurate in the long run.

      Delete