Changes in the use of literature with time - obsolescence revisited

One major study that might appear to be highly relevant is the University of Pittsburgh study (Bulick et al., 1976; Kent et al., 1979), which incidentally provoked a great deal of controversy, not least in the institution where it was conducted. This study collected, over a period of seven years, a great deal of data on library circulation and in-house use. One of the findings was that a high proportion (40 percent) of books acquired in 1969 had not circulated by 1975, and of those that had, nearly three-quarters were borrowed in the year of acquisition or the following year. However, the study does not demonstrate obsolescence so much as (1) an immediacy effect, and (2) selection decisions that proved to be inaccurate in terms of subsequent use.

A study reported by Hodowanec (1983) is of decay in the use of books by subject area in a university library. He found very large differences between subjects, foreign languages having the lowest use decay rate and business the highest. Rouse and Rouse (1979) also studied decay in the use of monographs, but by examining interlibrary loan demand rather than use in a library. They corrected the data for literature growth and duly found that this increased the median use age. They also found that demand at the regional level decayed faster than statewide demand.

Hindle (1979) advocated the use of Markov models of book obsolescence in libraries using two studies as examples. Sinha and Clelland (1976a, 1976b) presented a complex model for acquisition and disposal, claiming that data obtained in two scientific libraries confirmed that a negative exponential function fitted the relation between average use and age. Sandison (1977a, 1977b) disputed the methods and conclusions of both articles, was duly answered (Sinha & Clelland, 1977, 1978), and, in one case, made a further response (Sandison, 1979). At the other extreme to complex articles, a simple introduction to obsolescence and weeding by Lancaster (1988) should be mentioned.

A theoretical paper by Verhoeven (1986) seems to be a good example of pointless and inapplicable theory. His attempt to apply an expectation of life formula to collection management is not unlaudable in itself, but he looked at average life expectancy in a stationary population. The great variation between items makes averages quite useless, while in no library in the world, except a dead one, is the book population stationary.

Perhaps the most important issue amid all the technical discussion of models and statistics is the practical one. It is assumed that libraries want to make rational decisions, but can any library be expected to collect the necessary data and carry out the necessary analyses? And are they much better off if they do? There is little evidence that librarians have attempted to use any scientific approach to discarding; rather, they have employed whatever crude methods can be used quickly. Not only have librarians--rightly, though perhaps not for the best reasons--made no use of citation data; they have not generally attempted to gather much data on use in their own libraries, largely because of the time and effort involved. After all, they might argue, selection is far from an exact science; why should disposal be? One answer to that might be that just because there is a hit-and-miss approach about selection of new material, it is all the more desirable to compensate for at least the errors of commission by a firmly based disposal program.