College and Research Libraries WILLIAM E. MCGRATH and NORMA DURAND Classifying Courses in the University Catalog The authors contrast the university catalog and the card catalog and conclude that the university catalog is the best guide to the university's current scholarly interests. They urge that librarians study and clas- sify courses therein, such as books, using the Library of Congress or Dewey Decimal schemes so that specific class numbers are grouped by academic department and become substantial spans. The profile can be used as a selection tool, as a correlating tool between curricu- lum, circulation and publishing, and as a device to aid weeding and shelving. General and specific guidelines for classifying, including a method for resolving apparent duplication of courses in different de- partments are presented. Time and unit figures are tabulated. Spe- cific steps in classification and editing are described. THREE EARLIER PAPERS in this journaP referred briefly to a device for tabulat- ing data on books published, purchased, or circulated. This paper provides justi- fication for the device, suggests several uses, and explains in detail how it is constructed. I ustification Traditionally, librarians have regard- ed the card catalog and shelflist as the best guide to the scholarly interests of 1 William E. McGrath, "Determining and Allocating Book Funds for Current Domestic Buying," CRL, XXVIII (July 1967), 268-72; "Measuring Circula- tion According to Curriculum," CRL, XXIX ( Septem- ber 1968), 347-50; and Ralph C. Huntsinger and Gary R. Barber, "An Allocation Fom1Ula Derived from a Factor Analysis of Academic Departments," CRL, XXX (January 1969), 51-62. Mr. McGrath is Directo1· of Libraries, Miss Durand is Head Cataloger, Univer- sity of Southwestern Louisiana, Lafayette, Louisiana. the university. The card catalog and the book collection were constructed with requests, for the most part, by fac- ulty whose interests were usually spe- cialized and whose tenure was not al- ways permanent. Current book choices therefore did not always represent the fundamental current curriculum. Even now it is the assigned task of the faculty library representative to re- quest books which reflect the teaching mission of his department. His book choices are rarely questioned, nor need they be. What might be questioned is whether all the relevant books pub- lished in a given year are actually re- quested. Faculty turnover and special- ized interests disrupt continuity and therefore may actually prevent the li- brary's collection from fully reflecting departmental interest and need. Useful, then, would be a record which accu- rately reflects current departmental in- I 533 534 1 College & Research Libraries • November 1969 terests, has relative stability, and has a fairly high degree of continuity. This paper proposes that the best such record is the university catalog, not the card catalog. The card catalog reflects the cumulative scholarly interests of the curriculum; the university catalog re- flects the current, changing curriculum. Courses and whole programs are added or dropped. Whatever its shortcomings, however it changes, the university cata- log is as thumbed and studied as the card catalog. Like books, courses are titled, listed and printed. What better record? Why not recognize the univer- sity catalog, then, and use it as a selec- tion tool and guide to the collection? We could, if courses were classified by the Dewey Decimal or Library of Congress schemes, as are books in the card cata- log. The DC and LC schemes are instru- ments too valuable to use on books alone. Other forms of information in li- braries have been classified: abstracts by the Oxford decimal classification in For- estry Abstracts and Soils and Fertilizers, and abstracts on cards in many special libraries. Indeed, information scientists have put much thought into the prob- lems of the structure and classification of knowledge in recent years. Why not college courses then? Courses are uniquely relevant and central to the li- brary's purpose. A list of class numbers or groups of class numbers resulting from classification of the univer.sity cata- log can be regarded as a department profile and can be used in many ways. For example, by correlating the class numbers to those in the American Book Publishing Record, or the British Nation- al Bibliography, clerks could use them to preselect books for critical review and final selection by a librarian. The list of class numbers could also be used as a tool to analyze the relationship of cir- culated books to courses. Study of the university catalog is one excellent way to learn about the cur- riculum and is an unusual exercise for catalogers. University of Southwestern Louisiana catalogers stated, after classi- fying 1,346 courses, that they had gained ther~from a far better under- standing of the university program. It gave them a perspective they did not have before. Using both the DC and LC schemes for an unconventional ap- plication provided insights not possible otherwise. In addition to the applications sug- gested above, several more are listed below. Undoubtedly there are others. In each of the following, class numbers can be grouped under each department in the catalog, creating a tabulating frame- war k for correlation analysis, and other statistical studies. 1. Assessing the collection. Class num- bers for existing as well as new courses and departments can be com- pared to the shelflist or other classi- fied bibliographies to ascertain ade- quacy or lack of material. 2. As a buying guide. Currently pub- lished books with class numbers fall- ing into the same groups as those in the university catalog should be first choices for purchase. 3. MARC tapes as a selection tool. Class numbers on the MARC tapes could be searched and compared to the list of class numbers in the university cat- alog. Titles of LC books with call numbers that match the class num- bers would have purchasing priority. Important material outside the class number profile · would be reviewed and selected by traditional methods. 4. As a guide to .the collection. Teachers may use groups of numbers assigned to each course as a guide to relevant material in the stacks. 5. Circulation. The number of books cir- culated in each departmental cate- gory could be profitably correlated with the number of books in the existing collection. Classifying Courses in the University Catalog I 535 6. In-library use. The books used in the library could be correlated with those taken out of the library. 7. Weeding device. Books with class numbers not in the listed groups can be regarded as having low relevance and if they have no other value may be considered for discard. 8. As a shelving and storage aid. Books in the highly relevant categories can be shelved in optimum locations. Those not in relevant categories can be stored in less accessible places. So much for the argument. The fol- lowing discussion is intended for those who may be interested in applying the device to their own situation. The Classification Process The first step in the classification process is to assign numbers to individ- ual courses in the same manner that numbers are assigned to books. There is no need, however, to restrict classifica- tion to one number. If more than one number applies, each may be listed. At the University of Southwestern Louisi- ana nearly two numbers per course were assigned and entered in the mar- gin next to the descriptions in the cata- log. A 3"x 5" card was then prepared for each number assigned, including, in ad- dition to the number, the department and course number. Editing then re- duced the number of class numbers con- siderably, as can be seen in Table 1. Each cataloger reviewed specific num- bers assigned to each department, and found that ( 1) many class numbers were repeated from course to course and ( 2) a long list of specific class num- ,bers had accumulated with very few gaps between. The editing process sim- ply consisted of ( 1) dropping all but one each of the many repeated num- bers while listing on one 3"x 5" card all the courses having that one number, and ( 2) listing on another 3"x 5" card the first and last of the long list of class numbers, it being agreed that all of the numbers in between were included. Long, inclusive spans were thus created. For example: QA 303- 320 was given to calculus courses; QA 331-355, to com- plex variables; QA 331-360, to complex analysis; QA 371-381, to differential equations. Since these courses are all in the mathematics department, the four TABLE 1 NUMBER OF DEPARTMENTS, CoURsEs, AND CLAss NuMBERS AND TIME NEEDED TO CoMPLETE PROJECT Number of departments 64 Number of courses 1,346 Total classes 0 before editing Dewey 2,579 LC 2,383 315 463 320 cataloger / hours 240 cataloger/ hours 560 cataloger / hours Total classes 0 after editing Dewey LC Time required to assign numbers Time required to edit numbers Total time Total courses classified per hour 1,346 courses 560 1 /h = 2.4 courses per hr. cata oger ours Time to classify and edit one course 60 minutes . · 2 4 = 25 mmutes ( average) . courses 0 Individual numbers and spans of numbers 536 I College & Research Libraries • November 1969 short spans were made one long span, QA 303-381, and labeled Mathematics. The apparently nonexistent numbers, QA 321-330 and QA 361-370 are as- sumed to exist for the sake of closing the span. The edited class numbers were then arranged sequentially. An exam- ple follows: QA 135- 263 Math QA 264- 265 Computer Science QA 266 Math QA 268 Computer Science QA 269-699 Math (absorbing the example QA 303- 381) QA 801- 820 Physics QA 821-835 Civil Engineering The numbers were then grouped ac- cording to department: Computer Science: Q 300-380 QA 74- 76.8 QA 264-265 QA 268 z 699c--699.5 Math: HF 5691- 5716 QA 11 QA 39 QA 135--263 QA 266 QA 269-699 QC 851-999 TA 329-347 Physics: QA 801-820 Civil Engineering: QA 821-835 Specific Guidelines To coordinate the work of University of Southwestern Louisiana catalogers and to establish uniformity, the authors developed the following guidelines and instructions. 1. Single listings. A number assigned to more than one course within the same department need only be listed once. It is useful, however, to list these courses on one card for cross-refer- encing. 2. Spans. Specific numbers should be grouped together, whenever possible, to form spans. Thus, if 574 and 574.1 are separately assigned, then they can be grouped together to form the span, 574-574.1. Additional numbers can be added to make the span even greater. In constructing a span, build up from the specific to the general. 3. Specific numbers. Use specific num- bers whenever possible to assure that each department has its own group of numbers, while at the same time keeping the number of spans to a minimum by making them as long and as inclusive as possible. The more specific, the more accurate; but the more inclusive, the less cumbersome. (See also Guideline no. 6.) 4. Class numbers not in schedules. Spans may include numbers not spe- cifically listed in the LC or DC schedules. For example, the span 184-186 may be assumed to include 185.5, even though no such number is specifically scheduled. 5. Ending a number. A number in a span will end with the last number for that subject listed in the LC or DC schedule. For example, biogeog- raphy ends with 57 4.99 rather than 574.9, because 574.9 does not include all the subjects between 57 4.9 and 574.99. It is not necessary to stretch the number out further, e.g., 574.99999 ... because no such num- ber is listed in Dewey. The last num- ber listed for that subject is 57 4.99. 6. Survey Courses. If a survey course is offered in a department which offers specific courses in the same subject, do not assign any number. If no spe- cific courses are offered within the Classifying Courses in the University Catalog 1537 general subject of the survey course, then assign as specific a number as possible. For example, instead of QA 1-935 for a survey course in mathematics, use QA 36 (Encyclopedic Works, Textbooks, Compends, . etc.) And instead of Q E 1- 996 for a course in Orientation Geology, use QE 26-31, QE 41, an4 QE 61 7. Objectivity. When you discover a class number for a subject which you think ought to be included in a course description, but the descrip- tion does not actually include the sub- ject, do not assign that number. 8. Duplication. (When a course in De- partment B is assigned a number al- ready assigned to Department A.) All decisions, even when the factors seem complex, can be reduced to an either I or situation. The technique is to determine what are the vital fac- tors, to organize them in an either/ or manner, then test them with the in- dividual case. To help resolve the problem of duplication, Figure 1 shows a How chart based on the either I or logic. To enable librarians to make an esti- mate of the time needed to classify the college catalog, University of South- western Louisiana catalogers kept track of the time spent on their study. The figures are given in Table 1. Conclusions University of Southwestern Louisiana catalogers, after the project, offered several observations worth repeating. Many of the problems were the usual ones associated with classification and were already familiar, but a number of the observations required careful study. The most difficult was Duplication of Courses (Guideline no. 8). If one de- partment offered a course which was identical or nearly identical to a course offered by another department, which one should be assigned the relevant class number or group of class numbers? The Guideline could not resolve every con- flict. Several duplications were unre- solvable and were set aside for consulta- tion with departments or even the cur- riculum committee. The catalogers felt that, ideally, one cataloger should classify the entire cata- log. But the work load was too heavy for one cataloger and so was divided among four, each cataloger being as- signed specific departments. To help reconcile the inevitably divergent inter- pretations of courses and guidelines, one person, the head cataloger, performed one final overall editing. Several professors, especially in the science departments, devoted many hours interpreting course content and suggesting class numbers. On the whole, they were quite interested in the proj- ect and thought it worthwhile. The catalogers ran into another prob- lem-an old one-the poor course de- scriptions. Quality of these descriptions varies greatly. Some are too wordy or obscure, others are too general or cryp- tic. We decided that course descriptions would be interpreted quite literally. No numbers were assigned to a subject not explicit in that course description. This, of course, meant that many large blocks of class numbers would not appear in our final list, an inherent feature of the entire project. The catalogers admitted that Guide- line no. 7 was also difficult to follow. To list class numbers for subjects we think are important was and is a great temp- tation. Important though they may be, the university catalog was the authority. We could not, on our own authority, list these important subjects unless they were in the catalog. Yes 538 1 College & Research Libraries • November 1969 Yes Keep th e class no. in D ept. A Drop from Dep t . B Keep the class no. in Dept. B Drop from Dept. A FIGURE l. Decision flow chart. What to do when a classification number assigned to a course in Department B has already been assigned to a course in Department A. The number fits both courses equally well. Classifying Courses in the University Catalog I 539 This raises the question of whether ex- cluded numbers are irrelevant numbers. Since we are aware of the considerable subjectivity involved, we prefer to say that most books with class numbers among the listed numbers have a high relevance, and that books with class numbers not among the listed numbers have a lower relevance. We expect that, occasionally, some books outside the profile will be highly relevant and that some books within the profile will be highly irrelevant. This system makes no judgments about the merits or quality of each book. It simply says that based on a subject analysis of the course con- tent, and the resulting class number pro- file, every book, whether a very good one or a very bad one, has either a high or a low relevance to the curriculum. Naturally, to assure year-to-year rele- vance the list would need annual up- dating with numbers for new courses added and those for dropped courses deleted. Ideally, we might rate the relevance of books on an arbitrary scale ranging from 0.0 to 1.0. Books with class num- bers matching those in our list would be perfectly relevant and would have a rank of 1.0. A book with a class number not in our list-for example, QA 267- might be assigned a relevance of .5 be- cause, after all, QA 267 is mathematics and we do have a mathematics depart- ment. A book with a class number for a subject not assignable to any depart- ment might have a relevance of 0.0. Fi- nally, it may sometimes be practical to rank only the class numbers. Such a scale of relevance, for books or class numbers, could be the next develop- ment in a system of classified courses. ••