Breaking down census QuickCheck models

Nathanael,

I'm not sure where to begin and odds are slim that I will have time to work through all the issues this week between deadlines and a business trip. But, let's see how far we can get in an hour this morning.

The immediate stumbling block for me is a "language issue."In your second paragraph above ("fly-over level") you're using basic words in ways quite different from the way they are commonly used in history research and source citations.  I suspect you are using them as they are defined in the field in which you are trained, but different fields use words differently to meet their own needs. Among those who research history, the following applies:

• A "work" or "piece of work" is not an author, but the product the author produced.
• A "publication" is a work that is published—of which the most important identifier is the title, which you're including under "work." If I understand you correctly, you're using the word "publication" as shorthand for place and date of publication and name of publisher. If so, for clarity—and explicitness of instruction—the precise terms "place," "date," and "publisher" need to be used.
• A "source" is the whole item from which we take our information—not just a part of it—and the description of our "source" includes all the bulleted details above. Apparently, you're using the word "source" as shorthand for "source-of-our-source"; but where our source said it found its information is a quite different thing from the source we're actually using.

You also write, in that second paragraph:  "Type of the specific chapter/database/etc., ... is only included if it's independent of the main publication."  Your intent here is not clear to me, so let me try rewording to express the purpose of that data field:

1. Citations have two purposes: (a) to identify where we got our information; and (b) to provide sufficient detail that our readers--or ourselves at a later date--can understand exactly what it is we used from a quality-analysis standpoint.
2. With point b in mind, when we cite a title of something from a website, we recognize that the title could represent any number of things: an authored article, a database with abstracts, a set of images of original documents, a graph, a map, a forum discussion, etc.  Sometimes the title itself will state the nature of that titled item. Most times, it doesn't. Therefore, immediately after citing the title, we have a field in which we identify what kind of material this titled item represents.

From the perspective of 1 and 2 above, the phrase "only included if it's independent of the main publication" doesn't fit. If by "main publication" you mean the website itself, then we can't say that a database, article, etc. is "independent" of the website. It isn't. It's one item that is offered at the website and usually only at that website. Thus, it is not independent.

The discussion at EE 2.2, particularly "Rule 2," might clarify this for you.

Nathanael also asked about punctuation, saying:

Item is separated from Source by a semicolon; all other elements are comma-delimited.

Actually, "all other elements" aren't comma-delimited. Details for the act of publication appear in parentheses. The standard format for books is this:

(Place: Name of publisher, date/year of publication)

When citing websites, the place of publication is the URL. We don't cite a publisher because that will be the creator of the website that we've already cited. We do cite a date of posting or date of access. But those details go inside parentheses and use a combination of colon and comma.

The use of semicolons vs. commas, in citations, follows the same rules that we apply when writing narrative. In a nutshell, when we have items in a series and one or more items has sub-parts, we use semicolons to separate the major items and commas to separate the sub-parts.

Thus, in a citation that follows the list of 8 bullet points above, the first 7 elements deal with one thing: identifying the material we actually eyeballed. The 8th deals with something else: where our source said it got its information. There are two major parts to that citation model—two "layers" to the citation. The semicolon is the delimiter between those two major parts or layers. Within each part, there will typically be a combination of other punctuation marks: commas, parentheses, colons within the parentheses, quotation marks, etc.

EE covers this in much more detail at 2.64, particularly the subsection "Separation of Citation Elements." (10.12 also covers a specialized situation.) This broader issue is also indexed under "semicolons vs. commas" and cross-indexed as "commas vs. semicolons."

Nathanael also wrote:

I'm having difficulty fitting a number of QuickCheck census models provided in Evidence Explained to this the model. Here are three models I've been puzzling over. 1. and 2. are from Evidence Explained, pp. 269 and 271.

1. 1790 U.S. census, Beaufort District, South Carolina, p. 492 (penned), col. 3, line 26, Mary Odam; digital image, Ancestry.com (http://www.ancestry.com : accessed 2 February 2007); citing National Archives and Records Administration microfilm M637, roll 11.
2. "1800 United States Federal Census," database, Ancestry.com (http://www.ancestry.com : accessed 8 December 2006), entry for Underhill Strang, North Salem and York, Westchester County, New York.

Nathanael, different types of sources have to be cited in different ways, depending upon the kind of source, the details that are essential to capture for that kind of source, and the location of the material. Above, citations 1 and 2 are two radically different things (as explained by EE 6.10):

1. Cites the original census which we happened to view at a website.
2. Cites the database created by Ancestry with an abstract of census data created by Ancestry employees. We’re not citing the original census at all.

Therefore, there are differences in our format.

Source 1 requires a 3-layer citation:

• Layer 1 cites the original census that we are eyeballing as an image and it provides the details for the original that we can identify from what we, ourselves, are seeing. It does not follow the “database/website” pattern of the 8 bullet points above because we’re citing the original record.
• Layer 2 cites the database and website that delivered this image. Layer 2 follows the pattern of the 8 bullet points above. Specifically, it follows the first 6 bullet points. We do not have to cite the 7^th because we’ve already (in layer 1) identified what it is that we are using from this database.
• Layer 3 provides the source-of-the-source data that bullet 8 calls for.

Source 2 requires only 1-layer or 2-layer citation. This is a simple citation of a database entry or abstract—not the original document. What we are citing is entirely an Ancestry creation. We have one item by one creator who is also the publisher.  It follows the pattern of the bullet points you give above for citing databases at a website, specifically bullets 1-7. It does not include bullet point 8 because how-to-cite “source-of-the-source” data is not being demonstrated in this particular citation. Including source-of-the-source information, when it is given by our source, is always the best practice, as EE makes clear throughout (the index has 13 entries for it). However, space limitations on this 992-page book do not allow us to demonstrate every possible point in every sample citation.

Nathanael also wrote:

Here are three models I've been puzzling over. .... 3. is from the above-mentioned discussion thread. ...

3. "U.S., Civil War Draft Registrations Records, 1863–1865," database and images, Ancestry, 2011 (www.ancestry.com: accessed 10 October 2012), digital image, 3d Congressional District, Albert Clough, Ingham, Michigan; citing ARC Identifier 4213514, Consolidated Enrollment Lists, 1863–1865 (Civil War Union Draft Records); Records of the Provost Marshal General's Bureau (Civil War), Record Group 110; National Archives, Washington D.C.

Nathanael, when we are citing materials, we have options. This is covered in the "Fundamentals of Citation" chapter (Chapter 2) at 2.47–2.52 "Source List Arrangements." Becaues of the complexity of National Archives materials, the initial sections of Chapter 11 ("National Government Records") also nees to be studied for background.  EE 11.11 "Basic Formats: Online Images” specifically explains the differences that are puzzling your, with its two examples for citing online images of NARA records.

Generically speaking, when we use images from a database, we typically choose between an emphasis on the original record or the titled database. This lies at the heart of your question about the difference in format between Sources 1, 2, and 3.  Those differences involve both custom and complexity.

Censuses: Traditionally, censuses are cited by the year, then place, then specific data. In the second layer of the citation, we then cite the medium through which we used it (the microfilm, the website, or the original archives, etc.).  That is the simplest format  possible, given the nature of the record set—which is why it has become the “traditional way.”

National Archive Records (other than censuses): These are far more complicated. Most of these records are still accessible only at the archives and don’t involve a database at all. Some have been microfilmed but not digitized, and so we use and cite the film in a 2-layer citation: first the original, then the film ID. Some have been digitized—either from the film or from the original; these we cite using some form of a database citation.

As a general rule: If we have multiple items from one database—either database entries (as in your No. 2) or images (as in your No. 3)—we will typically choose to emphasize the database. By doing so, we have just one "master source" to create in our source list, rather than many different citations to the individual records. If we are entering our research data into a database of our own creation, we can call up that one master source and most of the fields will automatically populate; then we can use the "specific item" field to identify the exact item we are using--followed by a layer in which we cite the source-of-our-source.

Your No. 3 follows this format. If we preferred, we could emphasize the record instead, by reversing the arrangement of the layers. Doing so would give us this 3-layer citation:

4. Civil War draft registration card for Albert Clough, 3d Congressional District, Ingham Michigan; imaged in “U.S., Civil War Draft Registration Records, 1863–1865,” database and images, Ancestry (http://ancestry.com : accessed 10 October 2012); citing ARC Identifier 4213514, Consolidated Enrollment Lists, 1863–1865 (Civil War Union Draft Records); Records of the Provost Marshal General's Bureau (Civil War), Record Group 110; National Archives, Washington D.C.

Either approach would be acceptable from the standpoint of an adequately informative citation. However, example 3 would be the better than 4 from the standpoints of data entry in our own database and the creation of one single citation for our source list.

To sum up a couple of thousand words in one sentence: citation involves understanding (a) differences in records; (b) our standard options; and (c) the patterns for each. 

You've given our readers a good opportunity to mull these basics!