Sunday, June 7, 2009

Ancestry Quirks - Census Source Citations

With all of the emphasis on creating Source Citations in genealogy reports, books, periodicals and record databases, I had hoped that would be one of the industry leaders when it comes to providing satisfactory and Evidence! Explained quality source citations in their web pages and printouts.

Unfortunately, they are not. Most of the source citations I've seen are of poor quality for genealogists and do not come close to meeting Evidence! Explained standards. For example, here is the 1900 US Census record page for my great-grandfather, Frank W. Seaver:

The census citation below the family list says:

"Source Citation: Year: 1900; Census Place: Leominster, Worcester, Massachusetts; Roll: T623_692 Page: 16A; Enumeration District: 1645."

A proper source citation for this entry would be (created by me using an Evidence! Explained template - it's not that hard!):

"1900 U.S. Federal Census, Worcester County, Massachusetts, population schedule, Leominster town, Enumeration District 1645, Sheet 16, p. 264 (stamped), Dwelling 258, Family 371,. Frank W. Seaver; digital image, ( : accessed 6 June 2009), citing National Archives microfilm publication T623, Roll 692."

At least the barebones source citation provided by lists the microfilm publication and roll number, the ED and the page number. That seems, to me, to be the bare minimum information required.

However, for the 1900 US Census, when the user imports the historical record into their family tree database and attaches it to a person, some of the census source details are lost. Here is the citation on the "Source Information" page for this particular source:

It says only: "Year: 1900. Census Place: Leominster, Worcester, Massachusetts; Roll: ; Page: ; Enumeration district: ."

The name of the person, the microfilm publication, the roll number, the page number and ED number are omitted. This occurs on every web page or printout of this record.

1900 is not the only census database that this happens on. The 1870 US census source record is shown below:

The source citation for the 1870 US census record says:

"Year: 1870. Census Place: ; Roll: M593."

No mention of the census place, the person's name, the roll number or the page number in the citation.

I checked the 1850 through 1930 US census data. The 1870 and 1900 censuses are the only two that are missing the details of the barebones, but inadequate, source citation noted near the top of this post.

The information for the census place, person's name, microfilm series and roll number, page number and ED number are available - they are in the barebones citation on the Summary page (like the first screen above) for every census record. The information just needs to be programmed better to include all of the important information in an accepted and standard format.

I hope that fixes the 1870 and 1900 census problem immediately, and provides acceptable source citations on all of their databases in the near future.


amyrebba said...

Wow thanks Randy. I thought there should have been more information on the source in Ancestry. Hopefully they will take a look at this problem.

Elyse said...

I have always thought that Ancestry's source citations were not up to par. How sad!

Geolover said...

Randy, thank you for highlighting this problem with citations to US Federal Census enumerations.

In-between-the-lines, however, is the problem that what is attached to the tree person is not the Census Record, but's anonymous extract or 'notes' on *some* of the data in the actual record (usually omitting at least places of birth and occupation, always omitting the other useful data). In addition to the notes' omissions, they often include persons who are not enumerated in a given household (sometimes adding persons who are not even in adjacent households), re-arrange the order of enumeration, and incorporate the usual frequent misspellings and misreadings (saw one just yesterday where the noter entered "brother" when the enumeration said "boarder").

In short, these 'notes' are **not** "the Record," but are some unknown person's opinion as to what part of the record states, often together with some computer glitch may add or subtract.

Therefore, the item actually attached should be cited as an extract by Anonymous of a given Census enumeration (Micropublication, Roll, Page, Place, Household/Family #, head of household citation; repository, URL).

Ancestry should please stop calling these "the Record." Many creators of trees add these, evidently without ever looking at the actual image.

There are added problems when the "Record" recites an index entry in a multi-source compilation that does not specify which source the item is from.

Or when a database is not correctly described at all, and the extract (both index entry in search results and "Record") incorporates an item such as part of a date that is **not in** the source item.

My present "favorite" of the last type is entitled "Delaware Marriages, 1645-1899," described this way:

"Dodd, Jordan, Liahona Research, comp.. Delaware Marriages, 1645-1899 [database on-line]. Provo, UT, USA: The Generations Network, Inc., 2005. Original data: Index compiled from town and county marriage records in microfilm, microfiche, or book format located at the Family History Library in Salt Lake City, Utah by Liahona Research (P.O. Box 740, Orem, Utah 84059). Specific source information is listed with each entry."

The actual source is microfilm of a card file located in the Delaware Public Archives, Dover, DE, that may have been compiled as part of the WPA Historical Records Project. The card file does index some actual Marriage Records and Marriage Bonds (the database on does not note type of record at all). It also indexes conclusions drawn by unknown persons from wills, estate administration accounts, and selected County Deeds. The card file gives the sousrce of the conclusion; the extract from the card file does not. The card file gives a date of the record from which the conclusion was drawn. The extract of the item on gives the year-date of the **source record** from the card file, and usually adds "1 January" to complete the date. Hence, in the database there is a "record" that AS married MS 1 Jan 1729. The source record from the card file was AS's will (that gives the full name of MS's brother), which was probated in 1729. AS and MS married long before this date, and the "1 Jan" part was added by some intervening computer emendation. It looks plausible, but this database has an extraordinary number of marriages purportedly occuring on New Year's Day. This should be a signal to the genealogist that there is some other source record of the given year, but the description of the database would not give you much of a clue.

How on earth would one be able correctly to cite this sort of dreck? Oh, and there is a CD out there (Broderbund/LDS, I think) with the same extracted listings, but *not* the card file images.

This sort of horrorshow occurs in many sorts of databases on

Abba-Dad said...

I think it's time we stopped the mediocrity at and avoid calling these quirks from now on. This is the largest paid genea-site and probably the most expensive as well. They need to up their level of professionalism and stop treating us like idiots. What you highlight here is not a quick. It's just not trying hard enough. And it teaches new genealogists bad habits that will one day cause them a lot of frustration.

We run into these "quirks" on a daily basis, it seems like. I think it's time TGN did something about instead of adding another theme over at - don't you?

Cheryl Fleming Palmer said...

Great information, especially for the new genealogist, who would accept the source from Ancestry as perfect!

Professor Dru said...

Thanks for sharing, Randy, especially for the newbies.