College and Research Libraries MELISSA D. TREVVETI Characteristics of Interlibrary Loan Requests at the Library of Congress A study of interlibrary loan at the Library of Congress in 1976 showed that academic libraries were the most frequent borrowers , and requests were most often for materials in the humanities. The Library of- Congress re- ceived proportionally more requests for old items than academic libraries did , and it filled 54 percent of all requests it received. Distributions of language and place of publication of requested items , as well as the time required for handling requests, were also investigated. AT THE BEGINNING of 1976 the new Librar- ian of Congress , Daniel Boorstin, initiated a full-scale review of the library's policies, or- ganization, and goals. As part of this effort, the task force carrying out the review com- missioned several user studies. Two focused on interlibrary loan: one a sample survey of interlibrary loan borrowers conducted by mail; the other a study of the characteristics of interlibrary loan requests received by the Library of Congress. This paper presents the results of the second study, which had two primary goals : first, to provide a factual context for the more subjective information being solicited by means of the survey questionnaire; and, second, to furnish data about the needs of a specific user group in order to assist the Melissa D . Trevvett is assistant head, reference section , National Library Service for the Blind and Physically Handicapped , Library of Con- gress . The author gratefully acknowledges the as- sistance of Daniel Melnick , Congressional Re- search Service, Library of Congress , who de- signed the methodology for the study; David E. Trevvett, who performed the SPSS runs and helped generally with suggestions and criticism; Mary Lewin, Anne Hitchings , and Everett johnson, who helped with the document selection process; and Kathryn K. Blair, who served as a reference for many questions about interlibrary loan at the Library of Congress . 36 I task force in dealing with problems of col- lection development . The source of data for the study was the library's file of interlibrary loan requests for 1975, the most recent complete calendar year at the time of the study and, therefore, the most recent "dead" file. These requests were stored in folders within file cabinets; the folders were ordered alphabetically by the geographic origin of the request-either state or foreign country-then loosely al- phabetically by city. Since time and manpower were unavail- able to perform a random selection from the individual documents in the file, groups of documents were first selected; then the final sample of documents was chosen from these groups. Specifically, the file was considered to be a collection of 14-inch segments; a number was assigned to each segment, and 225 of these were selected by means of a random number table. These chosen seg- ments were removed from the files, and five requests from each packet (sufficient to give about a 3 percent sample) were selected by means of a random number table. The final sample consisted of 1,114 re- quests. From each request the following data were keypunched: type of library sub- mitting request, subject area, date of publi- cation, language of the item requested, number of days taken by the request to reach LC, whether or not the request was filled, the number of days taken by LC to fill the request, and, for cases in which the request was not filled, the reason for failure. Frequency distributions and cross-tabula- tions were produced using the Statistical Package for the Social Sciences (SPSS). The results of the study are discussed below, item by item. RESULTS Borrowers. Requests came from every type of eligible library. Academic and research libraries ac- counted for 63 percent of the sample; fed- eral libraries, 10 percent; other government libraries, 1 percent; special libraries, 5 per- cent; public libraries, 4 percent; and foreign libraries, 18 percent. Since any library other than a secondary or elementary school · library is eligible to borrow from the Library of Congress, one might expect that the above breakdown of borrowers by type of library would closely reflect the national borrowing population. There are, however, several factors, re- lated to the library's special position and regulations, which probably make this dis- tribution specific to LC. First, at the time of the survey, the Library of Congress se- verely restricted borrowing by publi.c librar- ies; thus, the proportion of public libraries is probably smaller than it would be without this restriction. Second, because the library once had a statutory obligation to lend to federal libraries, it continues to be the recipient of a large number of requests from them. Finally, many foreign libraries view the Library of Congress, the "national" li- brary, as the logical place to send all re- quests, particularly since LC provides re- questers with alternative locations when it cannot lend. Therefore, the proportion of foreign libraries may also be unusually high. Subject of Requests. Humanities materials were the most fre- quently requested , accounting for · 44 per- cent of the requests. Science materials were second (24 percent) and social sciences, third (19 percent). The remaining 13 percent consisted of LC classes A-General Works (1 percent); Interlibrary Loan Requests I 31 G-Geography, Folklore, Sports, etc. (2 percent); M-Music (3 percent); K-Law (2 percent); newspapers (4 percent); and man- uscripts (2 percent). The difference in the sum of the percents listed here and the total given is due to rounding. A cross-tabulation of the subject field of request by the type of library (table 1) showed a significant correlation. In particu- lar, federal libraries requested far fewer humanities materials than one would expect on the basis of the marginal distributions alone, while academic libraries borrowed correspondingly more in this area. Federal and special libraries were higher than aver- age in science requests, while academic li- braries were lower. Language and Place of Publication. The distribution of requests by language was essentially the same as that . for the sample of academic libraries that Thomson examined in her study. 1 Only the fraction of requests for Russian materials differed noticeably: 8 percent of the Library of Con- gress' requests were in this language as op- posed to 3 percent for Thomson's academic libraries (table 2). As would be expected, the distribution of places of publication was similar to the lan- guage distribution (table 3). Again, requests received by the Library of Congress closely resembled those received by academic li- braries except in the case of Russian mate- rials. Date of Publication. Table 4 shows the distribution of requests by date of publication. This distribution is distinctly different from the pattern that emerged from the studies compared by Ste- vens. 2 Table 5 was adapted from Stevens' article with the addition of figures from this study. Although the time divisions are not identical, it is still obvious that the Library of Congress request pattern is decidedly dif- ferent. As one would expect, . it receives more requests for old materials than the other libraries compared and fewer requests for very recent materials. Outcome of Requests. The analysis showed that the Library of Congress filled 54 percent of its requests. TABLE I CROSS-TABULATION OF SUBJECT OF REQUEST BY TYPE OF LIBRARY Subject Ac-ademic Humanities Number 359 Row Pet. 74.8 Col. Pet. 51.8 Social Number I29 Sciences Row Pet. 60.8 Col. Pet. 18.6 Science Number 94 Row Pet. 36.2 Col. Pet . 13.6 Other Number 111 Row Pet. 74.5 Col. Pet. 16.0 Total Number 693 Row Pet. 62.9 Col. Pet. IOO.O Chi Square = 270.4 with 15 degrttes of freedom Significance = 0.0 Number of missing observations = 13 Totals may not equal 100 percent because of rounding. Type of Library Other Federal Govt . Special 2 1 IO 0.4 0.2 2 . I 1.8 33.3 I6 .7 27 I 8 I2 .7 0.5 3.8 24.1 33.3 13.3 78 0 41 30.0 0.0 15.8 69.6 0 .0 68.3 5 1 1 3.4 0.7 0.7 4.5 33.3 1.7 112 3 60 10.2 0.3 5.4 IOO.O 100.0 100.0 ~ (] £.. ~ (JI:l (~;) G- ::x:J (~;) c;, (~;) ~ ~ Total ~ Public Foreign ;:3"' I8 90 480 t"' ~ 3.8 18.8 IOO.O ~ 45.0 46.6 43.6 ~- 6 41 212' ~ c;, 2.8 19·.3 100.0 . ._ 15.0· 21.2 19.3 t:l ;s 9 38 260 s: ~ 3.5 14.6 100.0 ~ <;:: 22.5 19.7 23.6 '- 7 24 149 1:0 "'"'-t 4.7 16.I 100.0 1;0 17.5 12.4 13.5 . 40 193 U01 3 .6 17.5 100.0 100.0 100.0 100.0 Interlibrary Loan Requests I 39 Tho mson's Sample of Acade mic Libraries • English German French Russian Spanish Italian Latin Japanese Polish Dutch Portuguese Swedish Hungarian Chinese Czechoslovakian Others less than .2% TABLE 2 LANGUAGE OF PUBLICATION OF REQUESTS (IN RANK ORDER) Pe rcent 67 .1 10.3 9.0 3.3 3.1 2.3 .7 .5 .4 .3 .3 .3 .3 .2 .2 Library of Congress English Russian German French Spanish . Italian Japanese Latin Danish Czechoslovakian Polish Bulgarian Chinese Korean Indonesian Others less than .4% • Abridged version of Appe ndix XI , p . ll9, in Interlibrary Loan Involving Academic Ubraries . Thomson's Sample of Academic Libraries• U.S . Gt. Britain Germany France USSR Italy India Spain Japan Netherlands Switzerland 1 Canada Belgium Poland Others less than . 7% TABLE 3 PLACE OF PUBLICATION OF REQUESTED MATERIAL (IN RANK ORDER) Percent Library of Congress 45.0 u.s. 12.2 Gt. Britain 8.6 USSR 7.3 Germany 3.4 France 2.3 Spain 1.3 Italy 1.2 Japan 1.1 Argentina 1.0 Netherlands .8 India .8 Denmark .7 Canada .7 Rumania & Bulgaria Others less than . 7% • Abridged version of Appendix XII , p . 120, in Inte rlibrary Loan In volving Academic Libraries. TABLE 4 PUBLICATION DATE OF REQUESTED MATERIAL Relative Adjusted Absolute Freq. Fre q . Freq. (Pet.) (Pet.) Pre-1700 4 0.4 0.4 1700-1799 15 1.3 1.4 1800-1899 165 14.8 14.9 1900-1949 420 37.7 38.0 19~59 100 9.0 9.1 1960--U9 162 14.5 14.7 197~72 123 11.0 11.1 1973--74 103 9.2 9.3 1975 12 1.1 1.1 Date Unknown 10 1.0 Missing Total 1,114 100.0 100.0 Valid Cases 1, 104 Missing Cases 10 Perce nt 64.0 8.4 8.1 6.7 4.6 1.7 .7 .4 .4 .4 .4 .4 .4 .4 .4 Percent 43.8 15.1 8.0 7.5 6.0 2. 1 1.9 1.2 1.0 .9 .9 .8 .7 .7 Cum . Freq. (Pet.) 0.4 1.7 16.7 54.7 63.8 78.4 89.6 98.9 100.0 100.0 40 I College & Research Libraries • ] anuary 1979 TABLE 5 CUMULATED PERCENT BY RECENCY OF MATERIAL Last Last Last Last Last Study 3 Years 10 Years 15 Years 70 Years 75 Years (Percent) (Percent) (Percent) (Percent) (Percent) Stevens' Article* Palmour 21 58 92 Taylor 17 49 94 Reynolds U . of Washington 19 54 Washington State Library 25 70 99 This Study Library of Congress 10 35 83 •stevens, "A Study of Interlibrary Loan," p.339. Citations from Stevens to studies compared in the table are: A Study of the Characteristics , Costs, and Magnitude of Interlibrary Loans in Academic Ubraries , comp. by Vernon E. Palmour and others, prepared for the Association of Research Libraries by Westat Research , Inc. (Westport, Conn. : Greenwood, 1972), p.39. David W. Taylor and others, An Ope rations Research Study of the Pacific Northwest Bibliographic Center (Washington, D.C.: U.S. Office of Education, 1972), p.25. Maryan E. Reynolds, A Study of Ubrary Network Alternatives for the State of Washington (Olympia, Wash. : Washington State Library, 1970), p.9. This rate can be compared to success rates ranging from 64 percent to 83 percent found for libraries compared in Stevens' article. 3 A cross-tabulation of the number of re- quests successfully filled by type of library showed that all types of libraries were about equally successful in obtaining loans (table 6), with special libraries somewhat less suc- cessful than the others. As shown in table 7, requests for humanities materials were filled slightly more often than requests in the other subject areas. In the case of the unfilled requests, three reasons accounted for almost all the failures: 1. Material was noncirculating (35 per- cent). Noncirculating categories at the Li- brary of Congress include rare materials, materials in poor condition, local history and genealogy, periodicals, unusually large size materials, and "in print" items. 2. Material was "not on shelf' or charged to a user (32 percent). "Not on shelf' is the library's designation for items in short-term inside use, and no charges are maintained on these items. Only 2 percent of the figure consists of items with known charges. Part of the other 30 percent may also be items with charges, but what proportion is un- known, because in most cases only the loca- tion is checked for the item; and no check of the charges is made. 3. Material was not owned (24 percent) . The second category-"not-on-shelf' materials-has been discussed by Goodrum, who analyzed the "not-on-shelf' problem at the Library of Congress as a whole. 4 Among the factors Goodrum cited as generally af- __ fecting the "not-on-shelf' rate, three in par- · ticular seem likely to have a dispropor- tionately high impact on this rate for books requested on interlibrary loan . First, the high volume of circulation re- sults in interlibrary loan requests being more difficult to fill than requests from other users. Congressional requests are placed on a waiting list if the book is un- available, and requests from readers using the reading rooms can be resubmitted fre- quently. It is more difficult, however, for interlibrary loan requests to be repeated easily and often. Second, the fact that a larger number of older books are requested through interli- brary loan causes retrieval problems. Whereas only 17 percent of the books re- quested by the Congressional Research Ser- vice and from the general reading rooms were published prior to 1950, more than 55 percent of the books requested on interli- brary loan fell into this category. Thus a greater proportion of the materials desired for interlibrary loan may have been ad- versely affected by heavy use and are thereby more difficult to retrieve. Third, as determined by a Loan Division study, a percentage of books requested for loan were in place on the shelves but were not pulled in response to requests for them. At the beginning of the study (August 1975) 33 percent of the books reported "not on TABLE 6 CROSS-TABULATION OF SUCCESS IN FILLING REQUEST BY TYPE OF LIBRARY Outcome Academic Not Filled Number 306 Row Pet. 60.0 Col. Pet. 43.9 Filled Number 391 Row Pet. 65.6 Col. Pet. 56.1 Total Number 697 Row Pet. 63.0 Col. Pet. 100.0 Chi Square = 6 .2 with 5 degrees of freedom Significance = 0.2849 Number of missing observations = 8 Federal 55 10.8 49.1 57 9.5 50.9 112 10.1 100.0 Type of Library Other . Govt. Special Public 2 35 19 0.4 6.9 3.7 66.7 58.3 46.3 1 25 22 0.2 4.2 3.7 33.3 41.7 53.7 3 60 41 0.3 5.4 3.7 100.0 100.0 100.0 TABLE 7 CROSS-TABULATION OF SUCCESS IN FILLING REQUEST BY SUBJECT OF REQUEST Subject Outcome Social Humanities Sciences Science Not Filled Number 192 111 133 Row Pet. 37.8 ·21.8 26.2 Col. Pet. 39.7 51.6 51.2 Filled Number 292 104 127 Row Pet. 48.6 17.3 21.1 Col. Pet. 60.3 48.4 48.8 Total Number 484 215 260 Row Pet. 43.6 19.4 23.5 Col. Pet. 100.0 100.0 100.0 Chi Square = 18.6 with 3 degrees of freedom Significance = 0.0036 Number of missing observations = 5 Total Foreign 93 510 18.2 100.0 48 .2 46.1 100 596 16.8 100.0 51.8 53.9 193 1106 17.5 100.0 100.0 100.0 .... ;:s Total ~ Other a:. 72 508 0"' 14.2 100.0 ~ 48.0 45.8 ~ 78 601 ~ 13.0 100.0 ~ ;:s 52.0 54.2 ::x;, t':l 150 1109 ..Q 13.5 100.0 ;:: t':l 100.0 100.0 "-> ~ ~ ~ 42 I College & Research Libraries • January 1979 shelf' were actually found in place when re- checked by Loan Division staff; by the end of the study (April 1976) this had declined to 18 percent. The difference between ac- tual "not on shelf' and reported "not on shelf' could have been caused by incorrect call numbers on the initial requests or by the time lapse between initial requests and the rechecking, as well as by the simple failure to respond to the initial requests. Although the three factors cited in Good- rum's report explain to some extent the "not-on-shelf' problem as it affects interli- brary loan, additional studies focusing on this area and on the other areas of unfilled requests could prove very useful to the li- brary. If, for example, an analysis of mate- rials requested but not owned showed any distinct patterns, then acquisition policies could be revised to encompass these areas. Time Involved. Two dates are recorded on practically every request slip: the date on which the borrowing library completed the request and the date on which the request was re- ceived by the Loan Division of the Library of Congress. If the request was filled, the request slip also contains the date on which it was completed. The analysis of the amount of time be- tween the completing of the request form and its receipt by the Loan Division showed that, for a library within the continental U.S., this time averaged five days. Requests from elsewhere in North America and from Hawaii took an average of six and one-half days, those from Europe an average of elev- en days, and those from Mrica and Asia an average of nine and one-half days. There were no requests from South America in the sample. (Although some European libraries sent all of their requests via air mail, others used surface mail. Because of this, the average for Europe was higher than that for Mrica and Asia, where all libraries used air mail.) These averages are broadly indicative of the transit time between the borrowing libraries and the Library of Congress. It should, however, be noted that these times may include additional nonmail time, for · example, the time between the date written on the interlibrary loan form by the request- ing library and the date on which the re- quest was actually mailed. The data permitted only one other analysis of time required in the process: the time required by the library to fill a re- quest. The results (table 8) show great var- iations in the amount of time various units needed to fill requests-from almost six days to over twenty-seven. The Loan Division reports that since this study was performed the · time required to fill requests for material from the special collections has decreased substantially. The addition of staff designated to handle inter- library loans to several of the special collec- tion divisions has reduced the processing time to three to five days. Retrieval time for items in remote storage had also decreased by 1977-but only slightly-to about two weeks. TABLE 8 PROCESSING TIME AT THE LIBRARY OF CONGRESS Request filled from general collections (i.e., by the Loan Division) Request filled from special collections Music Manuscripts Orientalia Serials Law Microforms Request fill~d from remote storage locations Average Number of Days• 5.9 27.4 14.4 8.9 9.3 12.1 15.5 17.0 *Two points about this average should be noted. First, it does not include the time required for the item to be charged out, wrapped, and mailed; all of this is done after the searcher has completed the request. Second, the average includes December cases, which are special . In December every request filled after the eleventh is held until the twenty-eighth for mailing, and only the latter date appears on the interlibrary loan form . The study reported here has provided a basic description of the characteristics, out- come, and processing time of loan requests received by the Library of Congress. As interlibrary loan service continues to evolve Interlibrary Loan Requests I 43 under the influence of task force and plan- ning office recommendations, this study can serve as a base for measuring the direction and magnitude of changes that occur. REFERENCES 1. Sarah Katharine Thomson, Interlibrary Loan Involving Academic Libraries , ACRL Mono- graph, no.32 (Chicago: American Library Assn., 1970). 2. Rolland E. Stevens, "A Study of Interlibrary Loan ," College & Research Libraries 35:336- 43 (Sept. 1974). 3. Ibid., p.340. The single pass success rates that were compared were: 64 percent-Thomson, Interlibrary Loan Involving Academic Libraries, p.~5. 71 percent-A Study of the Characteristics, Costs, and Magnitude of Interlibrary Loans in Academic Libraries , comp. by Vernon E . Palmour and others, prepared for the As- sociation of Research Libraries by Westat Research, Inc. (Westport , Conn.: Green- wood Pub. Co., 1972), p .43, 46 . 83 percent-Interlibrary Loan in New York State (New York: Nelson Associates, 1969), p.27 . 4. Charles A. Goodrum, "Change at the Library of Congress," Library of Congress Information Bulletin 36:698-700, 711-12, 725-26 (1977).