3
Social Science Information Studies (1980), 1 (53-62) 0 Buttcrworths 53 BOOK REVIEWS A. Kent, and others. Use of library materials: the University of Pittsburgh study. New York: Marcel Dekker, 1979. (Books in library and information series, v. 26). 272~. ISBN o 8247 6807 8 Circulation data have always fascinated librarian-administrators. In less statistically sophisticated times such data were innocently flourished and flaunted in annual reports to justify past policies; or, to support increased budgetary demands. A convention well enough understood to be harmless. Now, with computers to produce and manipulate large masses of machine-readable data this period of statistical innocence has ended. No bad thing perhaps? However, it may be less easy to pass judgment upon the manipulation of statistics when such quantities are viewed uncritically as adequate replacements for the direct study of aspects of human behaviour which gave rise to such statistics. In other words when figures are seen, without qualification, as objectified human behaviour. Despite such a doubt the primary assumptions underlying the work of most quantifiers are (a) that sufficiently large masses of time-series data may be reduced to trends and formulas accurately descriptive of causal library/material-use relationships and (b) that such relationships, perceived and isolated historically, are sufficiently stable and reliable indicators to be employed in decision making regarding the future. This report accepts such assumptions. In addition it is informed by the prevailing administrative philosophy which, in this country, is associated with the Atkinson Report. However, that American libraries are being subjected to constrictive forces similar to those operating elsewhere is hardly reflected in the supporting references. Of the seventy or so citations only two, by my count, are from non-American sources. A curious professional insularity given the nature of the problems examined. The work reports the findings of a detailed and carefully explained investigation into ‘the use and cost of books/monographs and journals in the libraries of the University of Pittsburgh’ based upon circulation data, much of it machine readable, covering the period 1968-1975. There have been few investigations so thorough or covering such a period of time. The research was undertaken (a) to improve acquisitions decisions (b) to establish ‘the point when materials should be purged from the collection or placed in low-cost, remote, storage facilities’ (c) to assess the ‘critical points at which alternatives to local ownership are economically feasible’. Numerous findings and practical impli- cations are presented for consideration. A few are listed below. ‘Only 56 per cent to 60 per cent of the books and monographs added to the collection in any one year ever circulate.. . it was found that any given book purchased had only a slightly better than one chance in two of ever being borrowed.’ ‘For books and monographs added to the collection in any one-year period, approximately 40 per cent are used only once; 8 per cent are used twice, 6 per cent are used three times.’ In general, usage of journals is low and is primarily ofcurrent ,journals. It is current journals that are browsed in the main. Principal journal users are students, particularly graduate students. Previous references guide reading. There is a high correlation between internally and externally borrowed books.

Use of library materials: the University of Pittsburgh study: A. Kent, and others. New York: Marcel Dekker, 1979. (Books in library and information series, V. 26). 272p ISBN o 8247

Embed Size (px)

Citation preview

Social Science Information Studies (1980), 1 (53-62)

0 Buttcrworths

53

BOOK REVIEWS

A. Kent, and others. Use of library materials: the University of Pittsburgh study. New York: Marcel Dekker, 1979. (Books in library and information series, v. 26). 272~. ISBN o 8247 6807 8

Circulation data have always fascinated librarian-administrators. In less statistically sophisticated times such data were innocently flourished and flaunted in annual reports to justify past policies; or, to support increased budgetary demands. A convention well enough understood to be harmless. Now, with computers to produce and manipulate large masses of machine-readable data this period of statistical innocence has ended. No bad thing perhaps? However, it may be less easy to pass judgment upon the manipulation of statistics when such quantities are viewed uncritically as adequate replacements for the direct study of aspects of human behaviour which gave rise to such statistics. In other words when figures are seen, without qualification, as objectified human behaviour. Despite such a doubt the primary assumptions underlying the work of most quantifiers are (a) that sufficiently large masses of time-series data may be reduced to trends and formulas accurately descriptive of causal library/material-use relationships and (b) that such relationships, perceived and isolated historically, are sufficiently stable and reliable indicators to be employed in decision making regarding the future. This report accepts such assumptions. In addition it is informed by the prevailing administrative philosophy which, in this country, is associated with the Atkinson Report. However, that American libraries are being subjected to constrictive forces similar to those operating elsewhere is hardly reflected in the supporting references. Of the seventy or so citations only two, by my count, are from non-American sources. A curious professional insularity given the nature of the problems examined.

The work reports the findings of a detailed and carefully explained investigation into ‘the use and cost of books/monographs and journals in the libraries of the University of Pittsburgh’ based upon circulation data, much of it machine readable, covering the period 1968-1975. There have been few investigations so thorough or covering such a period of time. The research was undertaken (a) to improve acquisitions decisions (b) to establish ‘the point when materials should be purged from the collection or placed in low-cost, remote, storage facilities’ (c) to assess the ‘critical points at which alternatives to local ownership are economically feasible’. Numerous findings and practical impli- cations are presented for consideration. A few are listed below.

‘Only 56 per cent to 60 per cent of the books and monographs added to the collection in any one year ever circulate.. . it was found that any given book purchased had only a slightly better than one chance in two of ever being borrowed.’

‘For books and monographs added to the collection in any one-year period, approximately 40 per cent are used only once; 8 per cent are used twice, 6 per cent are used three times.’

In general, usage of journals is low and is primarily ofcurrent ,journals. It is current journals that are browsed in the main. Principal journal users are students, particularly graduate students. Previous references guide reading.

There is a high correlation between internally and externally borrowed books.

54 Book Reviews

It is practicable to use acquisition dates as the prime determinant of material transfer to cheaper storage.

‘It seems safe enough to say that there is a portion of the collection that reasonably should be shared among several libraries.’

The suggested sampling techniques are applicable to other libraries.

The established trends and relationships have a predictive value.

Whether these findings are generalizable beyond the particular circumstances of the University of Pittsburgh has yet to be demonstrated. The same may be said of the usefulness of the various techniques for data gathering. However, it may be asserted with greater certainty that, as it stands, the exercise goes little beyond the boundaries of superficial description despite the masses of statistics. For example, it may be worth knowing that such a small proportion of stock is used, or borrowed. The question arises, what is to be done with this descriptive fact? Very little apparently. ‘The problem, of course, is that the techniques for predicting which books and monographs are likely to circulate o, I or 2 times do not currently exist’. Indeed. Similarly, it may be useful to think in terms of shared acquisitions. Here, ‘the principal questions are how to identify those books and monographs before acquisition and to identify the cost-effective equilibrium points where access rather than ownership makes sense’. There are no answers to these questions either.

The analysis of historical circulation figures alone cannot be expected to provide adequate guidance for future actions. Apart from unexpected differences between future and past, there is the obvious weakness that bare figures offer no explanations, they say nothing about why circumstances are as they are. At best they may suggest that certain areas, or problems, are worth further investigation. Usually with different techniques involving the questioning or observation of people. Little of this fundamental research was attempted in this study. The lack of such work raises awkward questions. For example, hardly anything is said about the acquisition system and its operation at the University of Pittsburgh. No explanation is offered for-the prevailing material selection patterns. Yet the presented figures must reflect decisions taken within that system. In other words it is at least plausible that the observed and measured use patterns are substantially artefacts of the acquisition procedures rather than reflections of inherent, stable, user behaviour. In which case the predictive claims associated with the findings are considerably diminished, even within Pittsburgh University. Clearly, the worth of the study would have been enhanced if the statistical findings and interpretations had been related to the causal practices. Understanding would be closer if answers could have been suggested to such questions as-who orders the unread material? Faculty members? Librarians? Individuals? In committee? Are selection policies inadequate? Is it necessary to ac-quire redundantly to ensure that basic material is obtained? Why is certain material not read? IS it enough to claim that unread means worthless? etc. Something of this questioning approach is to be found in the section dealing with journals, but is rarely in evidence elsewhere.

The terms benefit and cost-benefit are employed throughout most of the text as synonyms fbr borrowing or use. A possible usage which requires justification. Only in the chapter entitled ‘A cost benefit model of library operations’ is an attempt made to grasp thr definitional nettle represented by these terms. In this section benefit from the library is viewed as ‘the elf‘ect of book use on the augmented value of human capital’. The impression that WC are here dealing with measurable, concrete, expressions is strength- ened by the employment of‘curvcs purporting, fbr example, to describe the behaviour of cost and benefit over time. However, the undertaking should not be taken too seriously.

Book Reviews 55

As the author explains ‘one cannot be too sanguine about isolating the effect of book use because of its close identity with other explanatory variables. It may be, too, that book use fails to isolate many intangibles associated with the library. In place of book use.. . one could substitute the size of the university collection.. . conceivably, the entire library budget could be tried as a proxy variable.,’ Perhaps, too, the old-style librarians had not got it so wrong after all?

In summary an interesting report, objectively presented, but flawed by its uncritical reliance upon figures unsupported by explanatory studies of the operators and users of the library system.

Norman Roberts

A. Kent, and others. Use of library materials: the University of Pittsburgh study.

The seven chapters of this volume are based on a programme of research at the University of Pittsburgh financed in part by a grant from the National Science Foundation. The principal investigator was Professor Allen Kent, who explains in the Preface that in terms of the simple equation X - Y = Z (where X represents the in-house costs of furnishing library services, and Y represents the comparative costs of resource sharing, and Z the saving from resource sharing), the present study is an attempt at the measurement of X. The hypothesis tested is that it is not possible for any library to claim credibly that it is buying for future needs (which it cannot know), nor does it make sense for that same library to seek more and more funds to buy a greater proportion of published output in the expectation that its future hit rate will be better. All that happens, according to the hypothesis, is that an ever-increasing amount of material in the library is little or never used.

So the intent of this study was to develop measures for determining the extent to which library materials are used and the full cost of such use. It did not surprise the research team that since much of the material for research libraries is rarely or ever used, the cost of book use is very high. The question then arises: what should research libraries do?

First, they must constantly monitor the use of all books and journals in their collections, since one of the principal findings of this study was that only 56 per cent to 60 per cent of the books and monographs added in any one year ever circulated. It follows that a library could halve its bookfund if there were techniques for predicting which books are likely to circulate. Unhappily, as the study notes, such techniques do not currently exist. The study is also aware of the implications in this respect for publishers; but found that the kind of rationalization of print runs, formats and marketing which would serve the academic community best is at odds, of course, with the profit motive.

Second, the study postulates the reasonable notion that there is a portion of the collection that reasonably should be shared among several libraries; but points to the difficulties of identifying which books should be so regarded before acquisition and where the boundary of cost effectiveness lies between access versus ownership.

Third, the study suggests a re-examination of our bibliographic control procedures: is the traditional card catalogue effective enough, is it (as many have suspected) something of a barrier to greater use of stock? The path forward, therefore, is to grasp the new opportunities offered by more flexible and hospitable techniques such as machine- readable records.

Such studies as this volume pursues tend to bore, or daunt, librarians. It is about time though that our traditional attitude adjusted to the realities of present times. Accepting the logic of the situation that no library can acquire everything, no library can store everything, and no library can process everything, some new rules are going to have to be drawn up.