President Washington: Algorithm Can't Distinguish Non-sense from Good Sources

RobertTest

It is great to see that the Data Quality Score experiment is leading to some improvements in the profiles. However, it still isn't quite properly calibrated. It can't distinguish non-sense blather from well-sourced information.

I just looked at the profile on (President) George Washington KNDX-MKG born in 1732 in Woodstock, Connecticut—In Connecticut?-- Yes, according to the only formal citation provided for his date and place of birth. Fortunately, the documentation on his birth has been reduced from 30 sources to 6. Until a few months ago there were:

1. 17 References to Find a Grave or its clone Billion Graves;
2. 9 References to page 571 of the Roster of Revolutionary War Solders;
3. 6 References to an Iowa Grand Army of the Republic index card;
4. 4 References to a Netherland Genealogy index.
5. 5 other sources each referenced once.

So we have progress here.

Elizabeth Shown Mills in her 700 hundred page book "Evidence Explained", observes that "multiple sources for a particular statement confirm each other only when each is a reliable source of independent origin offering firsthand knowledge" (p. 21).

In the Details section of Washington's profile, under Birth data, two citations are provided that refer only to his Burial -- But this burial data is apparently considered important high quality birth data.

Of course there is only one firsthand source for Washington's birth -- his mother's Bible.
Or am I blind to the importance of the Connecticut birth records?

For some reason, under the Sources Tab Martha's Bible is not cited as a source in the current profile. The term citation has a specific meaning -- i.e., a specific set of criteria must be met for a statement to be an actual citation.

In the Sources section there is a specific spot where the citation is to be placed. There is no citation referencing his mother's Bible there. There is only a URL or a web link to an image of the Bible page. But it is the actual Bible that is the source that should be referenced and not a digital object. We can use the digital image but it has evidentiary value only because it is an image of an actual Bible page. It's that physical page that is the only important source. It must be cited.

Citing a source is an art: but every citation should include the title of the source - if it has one. If the source is an actual physical document such as a church record or a book or a page of a book then that document should be explicitly referenced. The citation would look something like:

Martha Washington's Bible, Identifier A-558, Washington Family Bible Page Mount Vernon's Manuscript and Special Collections Treasures database with images. (http://catalog.mountvernon.org/digital/collection/p16829coll5/id/631/) then the accessed date.
Original Source The George Washington Presidential Library Archives of the Mount Vernon Ladies' Association.

Now in Washington's profile there is a formal citation for his birth and his place of birth that does not refer to that Bible. That citation is

"Connecticut, Births and Christenings, 1649-1906", , FamilySearch (https://familysearch.org/ark:/61903/1:1:F74Z-J4S : 7 January 2020), George Washington, 1732.

It explicitly provides George Washington's vital data: He was born on 22 February 1732 in Woodstock, Windham, Connecticut, British Colonial America.

Washington's Birth Data in the Details Section rates a high quality score — generally free of contradictions and high in consistency. Accuracy? -- apparently too much for the algorithm to address at this point.

I don't mean for the above comments to sound snarky (they probably do) but I think George Washington deserves a lot better than this. As far as I can tell there is only one original firsthand source for his death — and the profile does not have a proper citation for it.

Here's the citation for his marriage:

Source number: 2750.000; Source type: Electronic Database; Number of Pages: 1; Submitter Code: PKS

The main purposes of a citation are to identify the information and provide the specific location where it can be found so that the reader can find it. This citation is a failure. I can't find the record.

So, for sourcing Birth, marriage and death — the profile on George Washington completely fails. It should have a Low Quality Score.

Julia Szent-Györgyi

Going a bit off-topic for this group, but I disagree (strongly) with the assertion that a link to an image doesn't qualify as a source citation.

The Sources list on a profile should provide just that: identification of the sources for each conclusion on the profile. These could be personal recollections, paper documents, physical objects (such as gravestones), or recordings of any of these, online or not. Most of them will be digitized or digital images, because none of us are travelling to Mount Vernon and gaining access to their special collections and actually looking at any Bibles. Or, more precisely, most of them will be links to digitized/digital images, because most of us are not in the business of hosting internet repositories.

It is my (very strong) opinion that such links are much more desirable than references to a unique physical object such as a family Bible. Even if the link is paywalled, it's useful to a far larger group of people than the location of the object. The ideal in my world is a link to the good-quality digital image on a non-paywalled site (such as FS), in a viewer in which one can page forward and back to gain full context. The identification of this image should include both the digital address (internet repository, breadcrumbs/cataloging info such as film number, etc.) and the physical/historical details (place, type of document or object, year, etc.).

I agree that that marriage "source" on George Washington's profile is a failure, as it doesn't actually identify anything — but I also think the index-only Connecticut source for his birth is likewise inadequate. In fact, I think the "no indexed sources" complaint has it Exactly Backwards: the scoring should be flagging profiles that have only indexed sources with no images available, because they're potentially perpetuating unverified data.

RobertTest

Hi Julia— thanks for the comment — I don't disagree that images are valuable and important.