Sometimes less is more. Recently the Mudd Manuscript Library addressed some long-standing problems with the Raymond Blaine Fosdick Papers to improve access to his voluminous correspondence (22 archival boxes, almost 10 linear feet). Fosdick, who is best remembered for his leadership roles in the League of Nations and at the Rockefeller Foundation, donated his papers to Princeton University in 1966. At some point, a portion of the correspondence in the Fosdick Papers was cataloged at the item level, meaning that (supposedly) there was a record of the author, date, and general subject matter of every single letter in that part of the collection. Each letter was (again, supposedly) also assigned a serial number, and the correspondence was arranged in numerical order according to these serial numbers. A database was available on an older version of the Mudd Library’s website that allowed researchers to do a keyword search of the item level descriptions, and the results would tell researchers the serial number(s) of the correspondence they might be interested in so they could request the relevant folder(s) through the collection’s finding aid. In the finding aid, however, the description of the correspondence just looked like this:

While the finding aid did refer researchers to the database, this method of providing access to the correspondence in the Raymond Fosdick Papers did not work well for several reasons. The most immediate problem was the descriptions in the database itself– and often, the lack thereof. A survey of the serial numbered correspondence revealed that a sizable portion of the letters in that part of the collection had actually never been cataloged, making these letters virtually impossible for researchers to discover. On top of this, the information about the letters that had been cataloged was full of errors and inconsistencies. Here’s an example of what the back-end of the database looked like:

In the greater scheme of things, this method was not ideal because it created an information silo– that is, it isolated the description of part of the collection away from the central method of providing access, via the finding aid. The existence of information silos make it more difficult for researchers to access archival resources, and having the information about resources in more than one location makes it harder for archivists to ensure that the information is accurate and consistent. To address these kinds of issues, archivists often incorporate the information from secondary access tools (in our case, the database) into their primary tools (the finding aid). For this particular collection, however, the descriptions in the secondary access system were in such bad shape that we decided to essentially start from scratch. Time for some good old-fashioned reprocessing!

We were still able to work with the information we had to give us a starting point for the reprocessing project. First, we did some cleanup of the names of the letters’ authors from the database– standardizing the format of names (e.g. the database included entries for “Baker, Newton” and “Baker, Newton D.”), fixing typos, etc. This gave us a long list of over 5,000 names, each associated with a different piece of correspondence. Next, we used an Excel formula to count the frequency of each name in order to generate a list of the correspondents that would probably warrant their own folder(s), rather than be sorted into a general alphabetic folder.

Then, a student worker did the time-consuming work of sorting all of the correspondence by author or subject, using the list of frequent correspondents as a guide. (Thanks, Zoe!) After making a few adjustments– combining some folders, creating new ones, and clarifying certain folder titles– the new inventory was ready to be added to the finding aid, and the need for the error-ridden database was eliminated. The end result looks like this (the entire reprocessed series is described here):

In addition to eliminating the need to cross-reference the finding aid with a database, having the correspondence in our finding aid system allows users to re-arrange the contents chronologically by clicking on the date column, if desired, and in any case, to request boxes directly, insuring an accurate page slip is created. Given the ongoing interest in Fosdick and this time period, we believe that these changes will significantly improve the accessibility of these important materials.