College and Research Libraries MIRIAM A. DRAKE Attribution of Library Costs Universities conduct a variety of cost-allocation studies that require the collection and analysis of library-cost data. Cost accounting meth- ods are used in most studies; however, costs are attributed to library user groups in a variety of ways. Cost accounting studies are reviewed and allocation methods discussed. Summary data from the Purdue University Libraries and Audio-Visual Center cost study are presented. uNIVERSITY ADMINISTRATORS conduct a variety of cost studies aimed at deter- mining the costs of instructional, re- search, and public service programs. The general purposes of these studies are to ( 1) examine the allocation . of resources; ( 2) provide information for future budgeting; ( 3) provide a basis for comparison of similar activities; and ( 4) provide data for government agencies and associations. While library services constitute a small portion of university budgets (usually 2- 5 percent), they are an essen- tial part of instructional and research programs. The costs associated with li- brary services, because they are joint or indirect costs, represent a difficult allo- cation problem for the cost accountant. This paper will review major library cost accounting studies, examine the methods currently in use to allocate li- brary costs to instructional programs, and present comparative data derived from the application of these methods. LIBRARY CosT STUDIES Cost studies conducted at four major university libraries-Stanford, Colum- bia, University of California, and Pur- due-are relevant to the topic of cost allocation. The Stanford study, con- Miriam A. Drake is assistant director, administrative services, Purdue University Libraries, West Lafayette, Indiana. 514 I ducted in 19641 and revised in 1970,2 was undertaken to determine the alloca- tion of library costs to four major user groups: undergraduates, graduate stu- dents, faculty, and staff. Cost elements were developed for each cost center or facility in the library system. Popula- tion and circulation data were gathered and applied to costs to determine the cost to be charged to each user group. The total costs for each group were di- vided by the population to yield per capita cost and cost ratios. A similar study was conducted by Gerald Quatman at Purdue University in 1961.3 The Quatman study was un- dertaken to provide data on the cost of providing library services for sponsored research programs. The study was unsat- isfactory for three reasons: ( 1) Stu- dents and faculty populations were not calculated on the same basis; ( 2) actual research use could not be measured di- rectly; and ( 3) the implicit assumption in the study was that "the average grad- uate student or faculty member engag- ing in sponsored research used the same amount of library services as the aver- age graduate student or faculty member employed in other university duties." In 1968-69 Columbia University con- ducted surveys of use and cost to de- termine the cost that should be allocated to research. 4 The allocation was based solely on use and not weighted by popu- lation. Analysis of survey data indicated that research accounted for 46.9 percent of library use; instruction 31.8 percent; both instruction and research 11.4 per- cent; and other 9 percent. The Leimkuhler and Cooper study, conducted at the University of Califor- nia in 1970, was undertaken to develop a cost-flow model for university librar- ies.5 Cost data were gathered and analyzed for cost centers, both process- ing and service. The study did not allo- cate costs to users; however, it did provide useful unit costs, such as cost per unit of circulation, cost per dollar acquisition, etc. In 1976 the Purdue University Librar- ies and Audio-Visual Center completed a cost allocation study that was conduct- ed for the purpose of supplying input data for the university's program cost study. 6 The specific objectives of the project were to determine the costs of providing library and audiovisual ser- vices and to allocate these costs to teach- ing departments and user groups within each department. The allocation meth- od_s used, and results, are presented be- low. ALLOCATION METHODS There are four basic methods for at- tributing the cost of library services to instructional and research programs. The first method uses faculty size as a base (usually on a full-time equivalent ( FTE) basis) and divides costs accord- ing to the proportion of faculty FTEs in each department. Student enrollment, number of courses offered, number of graduate students, and other variables that are related to library costs are not considered in this method. The underlying assumptions of this technique are that library usage for any single department or program is propor- tional to the size of faculty associated with the department and per capita usage is the same for all departments. Circulation data from a cost study of the Purdue University Libraries and Library Costs I 515 Audio-Visual Center show these assump- tions to be false. For example, the English faculty at Purdue represented 4 percent of the faculty population, but their library loans were 8 percent of loans to all faculty during the sample period. Per capita loans to faculty in the English department were 11.2, while per capita loans to the civil engineering faculty were 3.7. 7 The second method of library cost allocation is the standard technique for distribution of indirect costs and is based on faculty salaries. The National Center for Higher Education Manage- ment Systems ( NCHEMS) recommends that library costs be allocated according to compensation adjusted by faculty ac- tivity analysis. 8 This technique suggests that library usage varies in proportion to the compensation being paid and time spent on various activities. Higher salaried departments carry a higher share of library costs. In addition, the division of library costs between re- search and instruction, produced by the faculty activity analysis, is proportional to the time spent by faculty members in each activity. The application of this method does not account for the type of instruction or the type of research being performed by the departments. Humanities depart- ments, where research and instruction rely heavily on library resources, gen- erally provide lower faculty salaries than science and engineering depart- ments, where both teaching and research are more laboratory oriented and less dependent on the library. NCHEMS also suggests that a library can be costed directly if it is identified with a particular school or college. For example, a library that is associated with a college of business administration would be considered part of that col- lege, with the cost of library services allocated entirely to business adminis- tration programs. This method is based 516 I College & Research Libraries • November 1977 on intended use, which will be discussed below. METHODS BASED ON uSE The last two methods of library cost allocation are based on usage but differ in their assumptions and applications. One method is based on intended use or direct cost and the other on actual use. The intended use approach assumes that library materials and services are provided by a university with the intent that they be used by specific groups. While the technique has the advantages of direct costing and relatively simple data gathering, it does not result in an accurate summary of program cost. In earlier days when academic disciplines were purer, books purchased in a subject area were likely to be used almost ex- clusively by people in that discipline. With the increase in interdisciplinary studies and research, library materials in any one discipline are serving needs of people outside the discipline as well as people who are associated with the particular subject area. For example, the Purdue Libraries circulation study indicated that 58 per- cent of the loans made from the Kran- nert Library, School of Management, were made to faculty and students asso- ciated with programs in management or agricultural economics. People in fields other than those served by the library accounted for 42 percent of the loans and included engineers, home econo- mists, pharmacists, and others. 9 Trends in both research and instruc- tion suggest that interdisciplinary activ- ity is increasing. As a result of this in- crease, usage patterns in libraries will show more diverse populations using the literature in all fields. It is no longer realistic to assume that economics books will be used exclusively by economists or that use of psychology books will be restricted to psychology majors. A more accurate distribution of library costs will result if actual usage is considered. The actual use approach of allocating library costs is based on the theory that library services are offered to the total community and that the groups that use them should bear the burden in propor- tion to their actual usage. The method requires data on the identity of user groups and their use of library services. While statistics, such as number of reference questions answered and in- house use of library materials, are use- ful, they are costly to collect and process. Recorded circulation is the most convenient statistic to collect, and it gives a reasonably accurate estimate of the usage of the library by different user groups. The cost of collecting cir- culation data in libraries that use man- ual circulation systems will vary with the type of system being used. Hand tabulation of library use may not be cost justified. Allocation of library costs using the actual use approach, however, does provide a more accurate library cost component for university pro- grams. REsEARCH CosTs Satisfactory methods for allocating li- brary costs between research and instruc- tion in conventional academic libraries have not been developed. When faculty members or graduate students borrow materials from the library, the only way of determining how the material is to be used is to ask. This method was used in the Columbia and Stanford studies discussed above. The Columbia study concluded that 52.6 percent of library use was attrib- utable to research. The 1970 Stanford study allocated 30.9 percent of library costs to research. The primary drawback of user surveys to determine research or instructional use is that the user may or may not have a specific use in mind when library material is borrowed. When library users borrow books or journals, they may have some specific in- structional or research use for the ma- terial or they may borrow for back- ground or recreational reading. In addi- tion, library user surveys are expensive and time consuming for both staff and users. Universities that have large amounts of sponsored research generally use crude techniques for allocating library costs associated with research. One meth- od is based on faculty salaries adjusted by faculty activity ·analysis. Unfortu- nately, the method does not produce an accurate measure of the proportion of library resources being spent to support research programs. The method carries the disadvantages of the salary-based technique compounded by the inac- curacies in faculty activity analysis. PURDUE STUDY During the course of the 197 4-75 Purdue study, the allocation methods described above were discussed by the study team, librarians, and university administrators. The conclusion was that while the actual use method was most suited to the univer~ity' s needs, a com- parison of results obtained by the four methods would be useful. There was also extensive discussion of appropriate allocation techniques for research and instruction. The study team was unable to devise · a satisfactory method for di- viding costs between research and in- struction; therefore, total costs were allocated without regard to the end uses of library resources. The Purdue Libraries consist of a central processing department and twen- ty-four departmental libraries. The in- tended use computation is based on the costs of purchasing materials and the cost of operating service facilities in- tended to be used by a department or group of departments. The purchase of library materials at Purdue is controlled through fund numbers assigned to each teaching department. Purchase data were summarized for the fiscal year Library Costs I 517 1975 for each fund number. The aver- age cost of processing materials was add- ed to the purchase price to produce a total cost of materials for each depart- ment. The cost of operating the Physics Library was attributed wholly to the School of Science in the computation of intended use distribution. The actual use distribution consisted of ascertaining the total cost of operat- ing the libraries and actual usage of ma- terials by user group. The cost of operating the libraries was derived from fiscal and budget documents and includ- ed salaries and wages, fringe benefits, space and utilities, equipment, computer expense, and administrative, supplies, and miscellaneous expenses. Data on the identification of users of the libraries were not readily available; therefore, a survey of recorded circulation was un- dertake~ in the twenty-four libraries. The survey was conducted during the three busiest months of the year, Feb- ruary 1 through April 30, 1975. The re- sults of the survey provided the data: needed to allocate costs on the basis of usage. RESULTS Table 1 shows the allocation of li- brary costs to major user groups with comparative data from the 1961 Purdue study and two Stanford studies. The differences between Purdue and Stanford allocations are based, in part, on differences in relative enrollment and cost components. The higher percent- ages in the category of "other" for Purdue reflect the inclusion of staff in that group and greater use of the librar- ies by the public. Table 2 summarizes the distribution of costs to teaching departments ob- tained through the application of four allocation methods. The differences be- tween faculty FTE and faculty salary distribution generally are small. The greatest difference occurs in engineering where the faculty salary distribution is 518 I College & Research Libraries • November 1977 TABLE 1 DisTRIBUTION OF CosTs TO MAJOR GRouPs User Level Undergraduates Graduate Students Faculty! Other 0 Distribution of usage. 1961 Purdue 0 40.8 36.0 15.7 7.5 Percent of Total Cost 1965 1970 Stanford Stanford 15.4 43.7 36.4 4.5 24.6 37.4 33.8 4.2 1975 Purdue 43.9 36.7 9.5 9.9 t Stanford data include staff. Purdue staff included in other. 2.3 percent greater than the faculty FTE distribution. More significant differences occur in the comparison of cost distribution be- tween teaching faculty salaries and in- tended use. The intended use method indicates that the School of Human- ities, Social Science, and Education share of cost is 41.4 percent, compared with 24 percent using the method based on teaching faculty salaries. The share of the School of Agriculture indicates a 9.5 percent difference, and Engineer- ing shows an 8.8 percent difference when faculty salaries and intended use are compared. The distribution of faculty salaries is a function of both the num- ber and average salary of faculty mem- bers in the various schools. The average salary of teaching faculty in human- ities, social science, and education is lower than the average salary in agricul- ture or engineering; however, the pro- portion of library materials and facil- ities that are provided for humanities, social science, and education are far greater than those provided for agri- culture or engineering. The differences between the distribu- tions based on intended use and actual use are generally smaller than the differ- ences between faculty salaries and intended use. The greatest difference occurs in home economics in which in- tended use produces 1.9 percent share of costs and actual use indicates 8.9 per- cent. This difference reflects the smaller share of library budget that is allocated to the home economics library and ma- terials and the diversity and interdisci- plinary nature of the school's programs. The library materials needed by stu- TABLE 2 School Agriculture Engineering Home Economics Humanities, Social PuRDUE UNIVERSITY LIBRARIES AND Aumo-VISUAL CENTER DISTRIBUTION OF LIBRARIES' CosT BY ScHOOL, 1974-1975 Percent of Libraries' Cost Allocated by Teaching Teaching Faculty FTE Faculty Salaries Intended Use 17.5 16.4 6.9 17.1 19.4 11.7 3.1 2.7 1.9 Science, & Education 25.1 24.0 41.4 Management 3.7 4.3 8.6 Pharmacy 3.8 3.6 4.1 Science 20.8 22.0 17.9 Technology 5.5 4.2 3.0 Veterinary Medicine 3.4 3.4 4.5 TOTAL 100.0 100.0 100.0 Actual Use 9.8 14.1 8.9 38.7 6.2 1.9 16.3 3.0 1.1 100.0 dents and faculty in home economics are found in a variety of disciplines, in- cluding agriculture, creative arts, educa- tion, and management. CoNCLUSION The choice of a library cost . alloca- tion method will depend primarily on four factors: ( 1 ) the size of the library budget relative to the university's total budget, ( 2) diversity of instructional programs, ( 3) availability of data, and ( 4) philosophy of university manage- ment. In small colleges where educational programs are limited, allocations on the basis of faculty salaries or FTE's may be satisfactory. These allocations meth- Library Costs I 519 ods also would be acceptable in cases where the library budget is relatively small or management is willing to have some instructional programs subsidize others. In universities that offer a greater number and diversity of instructional programs and engage in sponsored re- search, usage-based allocation methods will give a more accurate accounting for the allocation of library costs. While the actual usage methodology is pre- ferred, it may not be possible to imple- ment because of the lack of usage data and the difficulty of data collection. When usage data are not available, · in- tended use will give a reasonably re- liable estimate. REFERENCES 1. Glen Densmore and Charles Bourne, A Cost Analysis and Utilization Study of the Stan- ford University Library System (SRI Project IM-5012; Stanford, Calif.: Stanford Univer- sity Libraries, 1965). 2. "A Cost and Utilization Analysis of the Stan- ford University Libraries System," unpub- lished, 1970. 3. Gerald L. Quatman, The Costs of Providing Library Services to Groups in the Purdue University Community-1961 (Lafayette, Ind.: Purdue University Libraries, 1962). 4. Ellis Mount and Paul Fasana, "An Ap- proach to the Measurement of Use and Cost of a Large Academic Research Library Sys- tem: A Report of a Study Done at Columbia University Libraries," College & Research Libraries 33:199-211 (May 1972). 5. Ferdinand Leimkuhler and Mic:hael D. Coo- per, Cost Accounting and Analysis for Uni- versity Libraries (Berkeley, Calif.: Office of the Vice President Planning and Analysis, University of California, 1970). 6. Miriam A. Drake, Libraries and Audio-Vis- ual Center Cost Allocation Study (RDU 76- 02; West Lafayette, Ind.: Purdue University Libraries and Audio-Visual Center, 1976). Available as ERIC document, ED 125566. 7. Miriam A. Drake, Circulation of Materials in Purdue University Libraries ( RDU 75-06; West Lafayette, Ind.: Purdue University Li- braries and Audio-Visual Center, 1975). 8. James R. Topping, Cost Analysis Manual (Field Review Edition, Technical Report no. 45; Boulder, Colo.: National Center for Higher Education Management Systems, 1973), p.159. 9. Drake, Circulation of Materials. NEW Up-To-Date Books From Noyes HARDCOVER BOOKS-LATE 1977 Data AUTOMOTIVE POLLUTION CONTROL CATALYSTS AND DEVICES by M. Sittig: A review of the affordable technology. The adoption of certain of these technical alternatives has occurred already and others will be introduced soon. Mandated fuel economy goals will require that fuel economy become a high priority design constraint. This is therefore an opportune time to survey the prac- tical aspects of catalysts, hardware and control devices. ISBN 0-8155~0618-3; $39 EPOXY RESINS AND PRODUCTS-RECENT ADVANCES by M. W. Ranney: Describes significant recent technological advances. The largest user of epoxy resins is perhaps the protective coatings industry. Nowhere else is the impact of ecological considerations more evident as powder, photo- curable, electrodeposited and high solids coating systems, using little organic solvent, are being commercialized rapidly. ISBN 0-8155-0679-1; $39 OFFSHORE AND UNDERGROUND POWER PLANTS edited by R. Noyes: Uncontroversial plant sites for generating electric power by any means are becoming increasingly scarce. Consequently, utility companies are looking out to sea or beneath the surface of the earth for suitable power plant locations. This is a most exhaustive and detailed treatise on this timely subject, based on government-sponsored reports. ISBN D-8155-068D-5; $42. TEXTURED PROTEIN PRODUCTS by M. Gutcho: There is a major effort in the world today .to pro- duce nutritious, high protein, meat-like products of other than animal origin, suitable for human consumption. They are made from vegetable proteins processed primarily by wet fiber spinning and thermoplastic extrusion. A large variety of new, innovative methods are described in this book. ISBN D-8155-0681-3; $39 YEASTS FOR FOOD AND OTHER PURPOSES by J. C. Johnson: Yeasts can proliferate without sun- light, yet they contain many valuable nutr_ients including proteins, fats, minerals and B-complex vitamins. The bulk of the single cell yeast protein is in the form of enzymes and ' is a valuable source of them or of nutritional single cell protein .' Over 160 production and use processes are described. ISBN D-8155-0682-1; $36 · PARTICIPATIVE MANAGEMENT-QUALITY OF WORKLIFE AND JOB ENRICHMENT edited by R. B. Miller: This is a book about management systems where everyone profits directly from in- creased productivity. Imparts detailed knowledge of the mechanics of such a system and the prob- lems likely to arise. ISBN Q-8155-0683-X; $18 RADIOACTIVE WASTE DISPOSAL-LOW AND HIGH LEVEL edited by W. R. Gilmore: Based on fed- erally funded studies, this book is concerned with the technological aspects of processing and storing both high level and low level radiqactive wastes (radwastes). Detailed accounts are given of radwaste storage and disposal at government-operated and commercial nuclear facilities. ISBN D-8155-0684-8; $39 PHYSICAL DISTRIBUTION MANAGEMENT -ANALYTICAL APPROACH TO CUTTING COSTS ·by R. Willis: Distribution of most merchandise requires freight transportation, warehousing, materials handling, protective packaging, inventory control, order processing, demand forecasting and cus- tomer service. This book provides simple analytical techniques that can be used gainfully by_ the modern distribution executive. ISBN 0-8155-0658-5; $18 ndc NOYES DATA CORPORATION NOYES BUILDING, PARK RIDGE, NEW JERSEY 07656