Statistics for cataloging are maintained for a variety of reasons. The library is interested in monitoring the total size of the collection and the total size of various components of the collection. The library also uses cataloging statistics to monitor workflow: since new titles generally take less time and effort to catalog than serial title changes and editing existing records generally takes less time and effort than creating new ones, it is useful to observe the kind and amount of such work being done.
These data are useful in planning the number and level of staff to request when opportunities for changes in staffing occur, while, data on the changing proportions of different components of the collection, such as the proportion of serials to monographs, are useful in developing the budget. In addition, one's bibliographic utility may charge different rates for different processes so these statistics are also useful in monitoring that budget. Cataloging statistics may also be used to supervise the progress of training of individual catalogers and, if used with care, may provide objective data for some aspects of performance evaluation.
Traditionally, cataloging statistics have been gathered and managed manually. Individual catalogers keep tallies of their cataloging activity which are then compiled into institutional tallies at regular intervals, usually monthly and annually. Depending on the size and organization of the library, there may be a variety of compilations at different levels of the organization between the individual and the library as a whole.
For example, Pullen's Catalog Department is organized into units whose supervisors find it useful to develop statistical compilations to aid in managing their units' workflow and training activities.
The development of spreadsheet programs for personal computers inspired in me the desire to automate this labor intensive statistics-keeping function and I volunteered to make this a personal project. Accomplishing this goal turned out to be far more difficult than I or my department head ever anticipated.
For one thing, the training available in the use of particular spreadsheet products is grossly inadequate. The exercises and examples used never address the situations encountered in libraries, often being too simplistic, not to mention business-oriented. (The same criticism can be made of almost any application software not developed specifically for libraries, including wordprocessing and database programs.) Anyone setting out to automate cataloging statistics using a commercial spreadsheet program must accept that they will have to train themselves.
A more critical difficulty, however, is the actual complexity of cataloging statistics. It would seem that the calculation of cataloging statistics is very basic mathematics. One adds things cataloged and subtracts things removed from the collection. One rarely needs to multiply or divide and higher math, let alone fancy statistical operations, is not needed.
The simplicity of the math is deceptive. Complications arise when one begins to categorize the things being counted: titles v. pieces; monographs v. serials; book format v. other formats; copy v. original; new titles v. added volumes and copies; government documents v. general collection; one cataloger v. another; this month v. last month.
Because of the different uses to which the statistics are put, it is not unusual for an item, being cataloged to be recorded in more than one way. At Pullen, even a simple monograph represents three different categories: a title, an item and the source of copy. A new periodical may represent a title and the source of copy, but no item, since we catalog from the first issue but do not count items until they are added to the correction in permanent form (e.g. bound or microfilmed). An added volume to a monographic set or serial represents an item but not a new title, and therefore no source of copy either. …