College and Research Libraries L,etters To the Editor: I want to thank you for your thoughtful and urgent editorial, "To Those Who Dedicate Their Lives." [C&RL, Sept. 1984] NANCY S. OSBORNE State University of New York at Oswego To the Editor: You state, "I believe we need to challenge our assumptions," and I agree heartily. [CM Editorial p. 247, C&RL Ouly 1984)] We also need to be careful of unwarranted assumptions. On page 285 of the same issue the concluding sentence of the Karr article ''The Changing Profile of University Library Directors 1966-1981'' is much too sweeping in its interpreta- tion of the data. Based on the evidence presented, one cannot dare to say that the previous directors of ARL libraries were genteel, scholarly, even dilettantish any more than one can say the present crop are career-minded managers, administrators and technicians. These descrip- tions are not related to the characteristics examined and are gratuitous. They weaken a clear cut piece of work. CAROLINE M. COUGHLIN Associate Director, University Library, Madison, New Jersey To the Editor: I write regarding an article inC & R L; "College Libraries: A Comparative Analysis in Terms of the ACRL Standards," by Ray L. Carpenter (42, Jan. 1981, 7-18). Some work in which I am engaged led me to this article on account of its closeness to my interests on several points; the same institutional types and the population based on the typology, the library operation budget and ,_9perating expenditures within the population, and the rela- tionship of that finance matter to the advancement in the ACRL Standards for College Li- braries (1975) of the "six percent 'rule'." What is wanted is a view of the empirical reality of ''college'' expenditures for their li- brary operations: How much, in fact-indexed to a particular major and standard expendi- ture figure (the Educational and General Account)-was spent. The idea is straightforward in appearance. But in trying to operationalize the idea there are a lot of big problems. It is difficult enough for most individual periods, much less over any extended span of periods. Not- withstanding, some semblance of a snapshot of behavior appears possible by approaching the problem as Carpenter did. He meshed an available data compilation made by a federal agency-the National Center for Education Statistics/Higher Education General Informa- tion Survey (NCES/HEGIS)-with a modification of the Carnegie classification, which in turn forms the frame of reference of the Standards for College Libraries (1975). The reason why this "window" of possibility was available is that the agency data is collected and displayed disaggregated, by individual institution; while at the same time the Carnegie classification departed from the frequent typology practice of merely specifying factors or characteristics on which to assign any particular institution to a category (a practice as im- peccable in theory as it is frustrating in application) by doing the assignments itself. Hence, virtually each institution in the global population of U.S. higher education institutions is 81 82 College & Research Libraries January 1985 categorized by the classifications. One simply selects the desired period, takes the NCES/HEGIS data set in one hand, the classification in the other, runs them together, ma- nipulates them with descriptive statistics, and there is the behavior of the type for the pe- riod. Or so it looks. Carpenter chose the 1977 NCES/HEGIS data, which half brackets the date of the Stan- dards for College Libraries (1975). He mentions data currentness as his reason. Actually, the Carnegie classification version used in the Standards (1975) was published in 1973, and the exact state of being of each of the member institutions dates back to 1970. Thus, seven years separate the classification data from the agency data. (A revised version of the Carne- gie typology was published in 1976, based on 1976 data: But then this was not the version the Standards (1975) used.) For this reason, Carpenter might just as well have chosen the 1969 or the 1971 data set (there is no 1970 data set). Greater closeness between the time of the data and the time of the typology formulation might have served better to deal with the problem that the study population is fluid. This flux on the part of the population is the result of institutional'' transition.'' But these considerations are among the methodological and procedural problems facing researchers in this area. It is not quite as pat as it looks at first glance. Any approach is subject to constraints and limitations which should be clear to the researcher and made clear to the reader. Hard choices, arbitrary choices, must be made. Carpenter made his. He is quite clear that, although he was working with the NCES/HEGIS 1977 data set, it was the machine readable form he used. It was not the paper copy Library Statistics of Col- leges and Universities, 1977. Institutional Data. (Terre Haute, Ind.: 1980), which is what I used in assessing his article. That difference may be important. In the paper copy of the 1977 data the pertinent section is Table 3. In the Introduction we are told that ''Information obtained in the survey and presented in this report relate to two different time periods. Data on collections, ... expenditures, ... pertain to the 1976-77 ac- ademic year. . . . '' (page 1; my emphasis). This statement is a potential source of confusion in that it may appear somehow related to a real problem that exists on the face of the data set. It is not, and may be disregarded. To see the real problem, look closely at any page in Table 3. The table is captioned "IN- DEXES CONCERNING OPERATING EXPENDITURES, 1976-77 .... "Beneath that, the collective caption of Columns 4-6 reads "LIBRARY OPERATING EXPENDITURES (IN DOLLARS) 1976-77." But now look atthe caption of Column6. It reads" AS PERCENT OF GENERAL EXPENDITURES, 1975-76." These readings seem to be consistent throughout Table 3. This is, on its face, a disparity. Taken~complete, what it says is "Here in Column 6 you have Library Expenditures for 1976-77 expressed as percent of General Expenditures for 1975-76." Put differently, it is only reasonable to interprete the literal meaning of these combined captions as saying that the numerator of the crucial index or percentage comes from one distinct time period, the denominator from another . . . the distinct time period next fol- lowing. It is 1975-76 + 1976-77. Is this disparity a misprint? Or is it a substantive, but one time, fluke in the paper copy of the data set? Suppose Carpenter had instead elected to more or less align in time the Stan- dards (1975) with the NCES/HEGIS data available for 1975. In paper copy form this would be Library Statistics of Colleges and Universities, 1975. Institutional Data. (Washington: GPO, 1977; NCES 77-206). There seems to be no 1976 data set. The tables are numbered differ- ently. But the pertinent one, again consistently, captions itself ''INDEXES CONCERNING OPERATING EXPENDITURES, 1974-75," while the column is captioned "AS PERCENT OF OPERATING EXPENDITURES, 1973-74." Again, the same principle of disparity. These disparities may be old stuff to some readers, but they surprised me. Suppose that Carpenter had instead chosen to half bracket the Standards (1975) on the other side. There having been no 197 4 data set published, the next closest would have been 1973. But in that year's paper copy form the ratio in question was not included in the data display. Suppose, instead, that Carpenter decided to publish an update, based on the principle of Letters 83 data currentness. In that event, he could avail himself of the 1979 data set, the paper copy of which is Library Statistics of Colleges and Universities, 1979. Institutional Data. [Springfield, Va.: NTIS; NCES 81-260]. (He would still be bound to the version of the ACRL Standards currently in force, the 1975 version, and thus to the 1973 Carnegie classification.) In the 1979 data set the pertinent table is number 3, which is captioned "INDEXES CONCERN- ING OPERATING EXPENDITURES. 1978-79 .... "The collective column caption reads "LIBRARY OPERATING EXPENDITURES (IN DOLLARS) 1978-79," and the pertinent column, number 6, is captioned "AS PERCENT OF GENERAL EXPENDITURES, 1978-79." The face disparity is gone. Now if my interpretation of the literal meaning of the evidence is correct, then what we and Carpenter were presented by NCES/HEGIS-and I emphasize the source of the problem-is a measure which is not merely invalid, improper, or illegitimate. It is simply meaningless. The only possible value for such a measure that I can conceive is that if it were the only measure available, and if it were consistently available over time, then as a lagged measure of sorts it might serve to reveal the shadow or echo of a trend; an indicator once removed. But that is not the case here, either with the measure itself or with Carpenter's use of it. As I have noted, I have been talking about the paper copy of the 1975 and 1977 data sets, and Carpenter clearly stated that he used the machine readable 1977 data set. The answer may lie here. However, if the machine readable data set is configured and structured in anything resembling the same way as, and contains the same data as, the paper copy-so as to allow the latter to be generated from it-then the disparity may reach back to there and compromise Carpenter's numbers anyway. At the very least, knowing what we know about the 1975 and 1977 paper copies, a rebuttable presumption to that effect is raised. HENRY GRUNDER Williamsburg, Virginia Midwest Library Service You won't find more personal attention .. . anywhere College and university librarians: We have what you're looking for . While Midwest utilizes state-of-the-art computer systems , we realize machines can't do it all. So when you want to place an order, ask a question, or discuss a problem, you can call direct on Midwest's toll-free WATS line and conduct business on a name- to-name basis with your personal customer service representative. It's the kind of attention that Midwest has provided to college and university libraries for 24 years . Midwest Library Service 11443 St. Charles Rock Road Bridgeton, MO 63044 Call toll-free (800) 325-8833 Missouri librarians call toll-free (800) 392-5024 Canadian librarians call collect (314) 739-3100 ACQUISITION PERSPECTIVES 2. Book House guarantees to order each book you request , with regular claiming to publishers . From major publishers on open account to the most obscure press requiring pre- payment and offering no discount, our persistence and experience enables us to deliver every available title. There is no substitute for complete delivery. CALL TOLL-FREE TODAY 1-800-248-1146 In Canada & Michigan CALL COLLECT (517) 849-2117 OCLC Vendor No. 17397 SAN 169-3859 the BSOK H IJSE JOBBERS SERVING LIBRARIES WITH ANY BOOK IN PRINT SINCE f962 208 WEST CHICAGO STREET JONESVILLE, MICHIGAN 49250