The Counts Project: Process and Methodology (2019-2020, Completed)

As part of my position as a Research Assistant on The Stainforth Library of Women’s Writing project, I was tasked with updating our understanding of the size of the library by way of counting the number of different editions that are in the Stainforth library catalog, which records the thousands of titles and author names in what we believe to be the largest private library of Anglophone women’s writing collected in the 19th century. The last attempt to quantify the size of Francis Stainforth’s library was completed in June 2016 by Cayla Eagon, before many editorial phases that refined the catalog transcriptions and database. A lot has changed since Summer 2016! We expected our count of the number of editions in the library catalog to drop, and it did.

Summary: There are 7122 editions in the Acquisitions section of the library catalog. The Wants section, or “wish list,” mentions 946 editions, a portion of which are also added to and counted with the Acquisitions section. These numbers do not include editions that were added and then deleted from the catalog.

Keep reading to learn about how we very painstakingly arrived at these numbers. 

At the outset of our 2019-2020 counting process, we defined what counts as an “edition” with the following terms:

  1. A title with only one publication date is one edition
  2. If there are several publication dates for a single title, each publication date counts as a separate edition
  3. Any unbound object, like a broadside or a clipping, counts as an edition
  4. References to collections or miscellanies usually refer to one text within that collection; that title counts as one edition

We acknowledge that our definition of “edition” is a bit loose. For example, of course there can be more than one edition published in the same year. But given the specific contents of Stainforth’s library catalog entries, the edition, considered in these terms, serves as a helpful, approximate unit of measure to quantify the number of items in the catalog, and by extrapolation, the extent of Stainforth’s library project.

After defining the terms of an “edition,” our next step was to establish and document the processes for counting them. The first time this count was done was at the beginning of the Stainforth Library research process; the catalog was in the process of being digitized, and because of this, editors counted by hand. Luckily, during this round of counts, Dr. Leuner was able to create an Excel spreadsheet of all of the listed titles in Stainforth’s catalog—both the Acquisitions and Wants sections of the catalog—and write a formula that automatically counted the number of publication dates associated with each title as a rough count of the number of editions. This saved me a lot of time counting those editions. However, this formula was unable to detect titles that have “No Date (ND)” in the catalog, lines with multiple editions listed, and any partially or differently written entries with publication dates.

The Acquisitions Section

I began by updating the counts of the Acquisitions section (pages 1-509). To combat the inabilities of the spreadsheet formula, my first steps were to: (1) go through the spreadsheet and, by hand, insert the number of editions for lines that the spreadsheet counted as “0” due to the reasons listed above; and (2) compare the “counted” editions in the spreadsheet with the lines of Stainforth’s catalog and adjust any miscounted lines. Although this seems simple and quick, the Acquisitions section alone is 509 pages with (at least) 24 lines per page. While I worked efficiently, I also worked carefully to acquire the most accurate count I could.

After getting an initial Acquisition count through these steps, I moved onto checking for duplicate titles within this section (as opposed to duplicates between the Acquisitions and Wants sections). In order to do this, I searched each title using the catalog search page on the project website. When searching a title through this page, the results list the (available) page and line reference, title, author, publication year, and publication place. Additionally, I could sort by author or title, which allowed me to visually compare any texts with the same title to see if they were published in the same year and where else in the catalog they were listed. I did this for each title in the Acquisitions section, adjusting the counts when necessary, to get as accurate a count as possible in this section.

While searching the titles in the catalog search page accounted for the majority of duplicate titles, it did not show titles that are mentioned with “Cross-Reference” or “See-Reference” entries because these do not show up in our search results (this is a feature we are working on incorporating in a future phase). These catalog entries occur in entries where Stainforth points to other places in the catalog; these entries end in “See [Author Name]” or “See [Title]”, hence the “See-Reference” entry label. Prior to the counts project, I worked on another project where I linked all of these entries to their reference point. Using this data, I was able to determine which of these entries duplicated an edition between the see-reference entry and its reference point. After collecting that data, I adjusted the duplicates by eliminating the count from the See-Reference. While many of these entries did include duplicates, some entries are the only ones to list a specific edition which resulted in this additional step to sort through which see-references duplicate editions and which do not. 

After following these steps and making adjustments in the spreadsheet, I counted 7122 editions in the Acquisitions section.

The Wants Section

The Wants section, while shorter than the Acquisitions section, was a lot more complex to count for three reasons. The first is because of the Wants section’s organization. This part of the catalog is Stainforth’s “wish list,” and he divided it into four parts: British Writers, American Writers, Plays Wanted, and Annuals Wanted. Additionally, sometimes Stainforth would catalog a book on his wish list in all of the applicable sections. For instance, a play written by an American writer may show up in the American Writers and Plays Wanted sections. Because of this, the counting of the Wants section required double-checking each edition across the four sections.

Second, while many editions were counted through the Excel spreadsheet by publication year, I had to hand-count editions with no publication year and adjust for lines with multiple editions. And third, reading the Wants section differed from the Acquisitions section due to the different types of strikethroughs throughout this section. Not all titles are crossed out, but lines may be having a single strikethrough, double strikethrough, or squiggly line through it. In cases where publication dates are listed, these dates may be struck out through single or double hashes. Because of the variety in strikethroughs, part of the process also entailed interpreting what each type of strikethrough meant in context with the editions. During this process, I added the additional step of noting in the Excel spreadsheet the type of strikethrough on any struck-out lines. We determined each type of cross-out to reflect the following:

  1. single straight-line strikethrough: this counts as an edition that Stainforth later acquires (many of these are in entries from the Acquisition section)
  2. squiggle strikethrough: counts as a deletion, not an edition
  3. double strikethrough:  counts as a deletion, not an edition
  4. hash marks: count as editions that Stainforth later acquires

This system is not without its exceptions. We have found cases of double strike-throughs that have entries in the Acquisitions section and single strikethroughs that do not. However, sustained patterns in the data led us to count the entire Wants section this way, and we were consistent throughout the count.

After determining the different interpretations of strikethroughs, we decided to obtain two counts of the Wants section. The first includes “deleted” editions and the second does not. Once I adjusted all uncounted lines and noted the varying strikethroughs in the Wants section, I checked across the section for duplicates. The Wants section count, after checking and adjusting for duplicates, is 999 editions. 

After coming up with this initial count, I wrote a formula in the excel spreadsheet to count the number of squiggly strikethroughs in the catalog: 23 editions. I also did this with double strikethroughs: 30 editions. To attain a count without duplicates or deletions, I subtracted these numbers from the initial count of 999 editions, which results in a count of 946 editions.

What Did We Learn?

Although both sections were interesting, this project really familiarized me with the Wants section of the catalog, which I had not worked with as closely before. By cataloging the different types of strikethroughs and investigating whether they were reflected in the Acquisitions section, I feel that I learned more about how Stainforth organized his catalog as well as gaining a general sense of the order in which he collected the editions. 

This section was a lot more complex than I, or Dr. Leuner, presumed, and analyzing the detail to which Stainforth cataloged his book collecting was fascinating. As I continued working with the data, there seemed to be more and more details that I wanted to track, and I was forced to limit myself for the sake of time. While the catalog, seems straightforward at first glance, a closer look reveals that it is extremely nuanced, which I completely did not expect. I look forward to continuing research and analysis on the way women writers and their work are cataloged in the Stainforth library.


You Might Also Like

One thought on “The Counts Project: Process and Methodology (2019-2020, Completed)

Leave a Reply

Your email address will not be published. Required fields are marked *