Towards a Theory of Fiction Replacement in Public Libraries
By Ken St. Andre, Librarian I, Phoenix Public Library
I have a dilemma. I need to keep the library’s fiction collection strong, but I don’t have much money for buying replacements for all the books being lost. I need to prioritize and only buy those items that are most needed.
I don’t need to worry about new books. Those decisions are being made for me (and the rest of the system) by Collection Development. We automatically get copies of new releases from all the best American authors and the best American publishers.
We don’t worry about small press items—no matter how good they may be. They don’t have the distribution to really affect the country.
In a public library, I am not really worried about the intellectual quality of the books I hand out—especially not in fiction. What counts is whether people want to read them or not. I leave considerations of quality to academic librarians working in university libraries. In the event that I ever have a patron who wishes to read something not found in my public library, I have the interlibrary loan system to fall back on—books may be obtained for them from the academic libraries that do order for quality instead of popularity.
My problem for replacement then is to determine which of the fiction titles are likely to be most popular, that is, most in demand. Fortunately, the library has statistical methods of determining where the true demand lies. There are two methods of measuring demand. One is the percentage of copies currently circulating. The other is the number of holds on the title.
By combining these two items, I should be able to determine the urgency of re-ordering any particular title once it goes missing.
One thing that must be considered is that I want titles for my own branch of any item that is even moderately popular within the city. Even though the branch is part of the system, and copies can be sent to where they are needed, it is qualitatively better to have copies here at my branch than to always be getting them from elsewhere. Thus if something is generally in demand, then I want at least one copy here.
I want a simple mathematical formula where a higher value is better than a lower value. I want factors for local demand, system demand, and future demand. Local demand would be how many copies does the branch have (C1) and how many are circulating (CC) with local demand being the percentage CC/C1. The maximum value would be 100%; the minimum would be 0.
The same formula should work for system demand. System copies would be S1; System circulation would be SC. Total demand would be SC/S1. Future demand could follow the same pattern where the number of system holds (H) would be divided by the number of available copies (S1). Adding the three factors should give a total demand factor.
A problem arises when the number of copies in the branch equals zero. Division by zero is impossible and produces infinity. I need a fudge factor to do away with the denominator ever being zero. For simplicity’s sake, let’s set that fudge factor to 1. Make it a given—a public library branch should always have at least one copy of any book that is in demand.
Having established my three factors, let’s make some rules for using them.
If branch demand is 100% then we need more copies of the book.
If System demand is less than branch demand, then we’re ok, because we can fill branch demand with books from other branches.
If future demand is greater than system demand, then the whole system needs more copies.
If the combined demand number is less than 100%, then no replacement copies need purchasing.
If combined demand is between 100% and 200%, then any branch at the 100% mark should buy at least one extra copy.
If combined demand is greater than 200% then more copies should be bought everywhere, and a larger branch should buy at least two copies.
Well, it looks good on paper. I wonder if such a theory could be tested.
Note that this model takes no account of such library realities as limited budgets. It also assumes a snapshot in time. Let's say demand is tested once every three months. A quarter seems like a reasonable period--it allows time for items purchased because of the previous analysis to be received and entered into the system, changing the branch and the system numbers. It allows time for holds to be filled and new holds to be placed. If demand still remains greater than 100 or 200 percent after a three month period, then additional copies should, in my humble opinion, be purchased again.
With the complexity of today's automated circulation systems, it should be possible to derive my various numbers from the computer without having a human being actually take the time to tally these up by hand. C1 and S1, CC and SC, and Holds are known quantities in modern computerized circulaton systems. All it really takes is some kind of reporting mechanism to extract these numbers, do the simple calculations, and prepare a report for the ordering librarian to authorize the purchase of replacement copies.
Nothing this mathematically determined will ever happen, of course. There will never be that much science in library science.