College and Research Libraries VIRGIL F. MASSMAN and DAVID R. OLSON Book Selection: A National Plan for Sinall Acadeinic Libraries -ALTHOUGH BUILDING THE COLLECTIONS is one of the most important tasks of li- brarians, comparatively little attention has been given to this aspect of profes- sional work. The system in current use has been practiced for many years with little systematic scrutiny and with little discussion of possible alternatives. This applies to nearly all academic libraries, but the present article will address itself mainly to problems of the smaller insti- tutions rather than those of the major university libraries. How are books selected for academic libraries? While patterns vary from one extreme to the other, in most institu- tions both librarians and faculty mem- bers participate in building collections. Many problems arise, for while faculty members play a major role in selection, librarians know very well that faculty selection is often of questionable merit. Among the most conspicuous deficien- cies are: ( 1) many faculty members are already overburdened with other duties; ( 2) some of them lack acquaintance with the world of books; ( 3) some do not care ( the textbook is enough) ; ( 4) a few suffer from a constitutional incli- nation toward laziness; ( 5) some select books in their own narrow field of spe- cialization without regard for the needs of students; and ( 6) some believe that only they are capable of selecting. While this list could be extended, these are some of the major shortcomings of Mr. Massman is director of libraries and Mr. Olson is head of public services at the University of South Dakota. reliance on faculty selection. As Danton has pointed out, the faculty member who fails to find a particular item in the library blames not himself or anoth- er faculty member for the deficiency, but the library for failing to procure the wanted title. 1 How about librarians? Certainly many of the problems which apply to faculty selection also apply to librarians -lack of time, inadequate acquaint- ance with books, and laziness. Librari- ans, however, usually maintain that they are more likely to consider the needs of students, and that they are more con- cerned about building a balanced collec- tion. Given an ideal bah~nce between selec- tion by faculty members and librarians, one might expect to develop a reason- ably good collection. However, because of the complexities of assembling a complementary library staff and faculty and of maintaining completely harmo- nious relationships between the two groups, this hope is a virtual impossibili- ty. ·Under present conditions the quality of selection in most academic libraries probably leaves much to be desired, but this is not entirely the fault of either the faculty members or the librarians, or even the two in combination. Why? Part of the defect results from the manner in which books get into review- ing journals. This itself has received comparatively little detailed study. The Bowker Annual lists the total number of titles examined in a number of gen- eral reviewing journals, but of course I 271 272 I College & Research Libraries • July 1971 makes no effort to assess the quality of reviewing nor the duplication of cover- age (i.e., whether a particular title re- ceived notice in more than one jour- nal).2 To a considerable extent the edi- tors of the reviewing journal depend upon the publisher to send new works for examination. The editor must then determine whether a particular book is suitable for review in his journal and give the book to a reader who may or may not return his . evaluation within the specified period of time. The latter situation is an especially vexing problem regarding reviews in scholarly journals. The specialist to whom the book is sent for examination is often busy with more pressing tasks, and may take six months, a year, or more to read the book, write his commentary, and submit it for publication-if he gets it done at all. That the current system is haphazard can be illustrated to some extent on the basis of difficulties encountered by CHOICE. This journal farms out re- viewing duties to a large number of li- brarians and faculty members, and the editor himself does not know what will be in each successive issue until virtual- ly the last minute. 3 What .appears in each number depends upon copy sub- mitted by reviewers. If the reviewer is dilatory, it may take him three months or six months to send in his report. This is not to blame the editor of CHOICE, for he is at the mercy of his geographically dispersed staff of volun- tary contributors. To manage such a task must take an unusual measure of patience and dedication. Nevertheless, even when it works well, the system leaves much to be desired. Thus, before the librarian has a chance to see the review and before a book receives a printed notice, the book must ordinarily be sent out by the publisher, meet the editor's standards, .and. await evaluation by the critic se- lected to review the work. The book re- view editor himself may reject many items, not because he necessarily ques- tions their merit but because the title does not fall into the subject categories or the type of literature (e.g., scholarly or popular) deemed appropriate for that journal. Because of the way the sys- tem works a large mass of literature, then, never comes to the librarian's at- tention unless he consults a large num- ber of reviewing journals. How unpredictable the vagaries of re- viewing journals are can be illustrated by taking five specific examples. Each of the five titles to be discussed was checked against the Book Review Digest and the Book Review Index to locate re- views.4 The first two examples are sig- nificant titles partly because they are of interest to minority groups. The other three are of value because they deal with certain aspects of higher educa- tion. All five books belong in every aca- demic library in the United States. Which journals reviewed these five books? The first example, published in 1967, was Donald C. Dickinson's Rio- Bibliography of Langston Hughes. Be- cause it contains extensive information about one of America's great black poets it is .a basic study which is essen- tial for anyone who is interested in the broad sweep of American literature, yet it received a notice only in Nation and Library Journal. The second work, Vine Deloria's Custer Died for Your Sins: An Indian Manifesto, was published in 1969 and was reviewed in America, Best Sellers, Library Journal, New Y ark Times Book Review, Newsweek, Satur- day Review, and Time. 5 The other three books chosen as ex- amples deal with issues that are of pri- mary concern to librarians and faculty members. The first, T. Caplow and R. J. McGee's The Academic Marketplace, discusses hiring practices of institutions of higher learning. It was published in 1958 and was reviewed in Library Jour- nal and the Chicago Sunday Tribune. The second book, Mark Ingraham's The Outer Fringe: Faculty Benefits Other than Annuities and Insurance, was pub- lished in 1965 and was reviewed in Ca- nadian Forum, I ournal of Higher Edu- cation, Library Quarterly, Science, and Teachers College Record. In this case, as is true for the next item, the subtitle gives a good indication of the content. Mark Ingraham's The Mirror of Brass: The Compensation and Working Condi- tions of College and University Admin- istrators was reviewed by CHOICE and byCRL. As already suggested these five items should be available in all college and university libraries, but no single review- ing journal covered all of them. Library I oumal noted The Academic Market- place but not The Outer Fringe or The Mirror of Brass. Of the three books just mentioned, CHOICE reviewed only the latter. ( CHOIC£ was, of course, not yet in existence when The Academic Mar- ketplace was published.) Similarly, the I ournal of Higher Education and Teachers College Record reviewed The Outer Fringe but not The Academic Marketplace or The Mirror of Brass. Oddly enough, none of the five journals which reviewed The Outer Fringe re- viewed The Mirror of Brass or The Ac- ademic Marketplace. Is there any ration- ale for this, or does it reflect the hazard of chance by which books are reviewed by one or another journal?6 Take a half hour to examine the Book Review Digest or the Book Review Index and see how many books which are of value to academic libraries are reviewed only by scholarly journals or only by the general journals. Further- more, see how many books which are of value to academic libraries are cited with only one review in Book Review Index. (The Book Review Digest nor- mally does not cite titles which received only one review. ) If one depends upon reviewing journals as a major source of information for building collections, Book Selection I 213 such an examination may be both en- lightening and disturbing to the person concerned about quality selection. A recent article in C RL discussed the reviews of books in seventy-one scholar- ly journals.7 Of the 3,195 titles exam- ined for that article, only about 15 per- cent received a notice in more than one of the ·seventy-one periodicals. Thus, 85 percent were reviewed by only one jour- nal. This meager duplication is rather surprising. One would expect far more overlapping within the journals for his- tory or for English, for example, or for any other discipline. Because there is not, however, it is necessary to examine at least several journals for each disci- pline, and the total number could easily come to seventy-five or more for all the various courses offered in the undergrad- uate curriculum in most colleges and universities. Even such extensive exam- inations of reviewing journals still would not assure the appropriate range of coverage-to say nothing about the quality of reviewing. In discussing the advantages and shortcomings of blanket order plans, comparatively little attention has been paid to the deficiencies of the current system of reviewing new books. A blan- ket order plan that is handled by a good dealer is probably capable of giving the library more effective coverage of cur- rent books than a system of relying up- on reviews. The study of the feasibility of cen- tralized processing in Colorado academ- ic libraries, for example, found that the approval dealer supplied 40.4 percent of the titles reviewed in CHOICE dur- ing the first year and 45.1 percent for the second year. 8 The writers suggested ~ that the approval plan needed to im- prove its coverage because it provided such a small percentage of the CHOICE titles. When the list of 3,195 books exam- ined for the C RL article mentioned previously and which received favorable 274 I College & Research Libraries • July 1971 reviews in the journals was compared with CHOICE, duplication approached only 30 percent. Thus the approval deal- er mentioned in the preceding para- graph achieved a higher overlap with CHOICE than CHOICE did with the scholarly journals. When the 3,195 titles were compared with the Book Review Digest, duplication approached 50 per- cent.9 Then there are also the general re- viewing journals such as Saturday Re- view, the New York Times Book Re- view, etc. In many instances, as an exam- ination of the Book Review Index will demonstrate, a book which is significant to academic libraries may be reviewed by only a scholarly journal or by only general journals. Thus to insure effec- tive selection, the library must devise a scheme which will assure adequate selec- tion based on thorough and r Jgular ex- amination of the general reviewing or- gans as well as a large number of spe- cialized journals which carry reviews. Under the present system it is ex- tremely difficult to insure the building of first-rate collections. It is, therefore, unfair to place undue blame on librari- ans for deficiencies in building collec- tions, for the present method is virtual- ly impossible to cope with. The librari- an may be doing an excellent job of se- lecting from those · journals which he finds time to read; it is impossible to read them all. To then use the standard procedure of evaluating the collection by checking it against recommended book lists and blaming the librarian if the collection appears to be deficient is affixing blame on a potentially innocent party. Much of the blame might more deservedly rest with the inadequate re- viewing system. · A new approach must be found. Al- though many librarians will object to .any suggestion of centralized selection, they should be aware of the fact that publishers and journal editors do a great deal of selecting simply by deciding which works will or will not be re- viewed. Over this the librarian has no control. For a moment, it may be worth ex- amining a few of the major objections to centralized selection. Presumably, the librarian knows his clientele, buys with individual readers in mind, understands their special needs, and is aware of how his people use books. But is there really any documented evidence that librarians (individually or en masse) know their communities as well as they think they do? What constitutes knowing the pa- trons? Does the opinion of one vocifer- ous faculty member speak for the fac- ulty? To turn to a slightly different area, librarians have strong feelings about whether sets should be classified as sets or whether journals should be classified with books. Is there any sub- stantial evidence that either sets or jour- nals are used more effectively in one way or the other? Is there any "scientific" evidence, in other words, which goes be- yond the unsubstantiated assumptions to which we cling so dearly, but which are at opposite points of the issue? Pos- sibly the arguments for local book selec- tion are similar. Possibly the librarian believes he can select more effectively for his patrons than anyone else, but he has no concrete evidence to support his view. Do most librarians select with some shadow of their own image (or the projection of themselves) in mind? Is it possible for the librarian to know what the vast majority of faculty mem- bers and students need? The librarian may heed the few whom he knows, but then he should be willing to admit he is doing that and nothing more. Much of the same holds true for selecting for subdivisions of the curriculum. The general content of American history is known; the facts are the same regardless of where American history is taught. One professor may stress the Civil War or immigration, but what if he leaves or if his course is dropped from the f curriculum? It does happen. Further- more, if a professor or if the entire faculty stresses a particular aspect of American history, the library still needs the important works dealing with other aspects of that subject. Carried to its logical extreme, the con- cept of selection for present clientele would necessitate the reorientation of the collection every fall when the new crop of students and faculty members arrives. And what happens when this li- brarian with his extraordinary insight into the needs of his clients leaves? Will he then be capable of immediately ad- justing his extrasensory wave lengths to his new clients at another institution? The contention that the librarian is se- lecting for particular individuals sounds convincing. However, if the librarian is indeed buying particular titles with the needs of one person in mind, is he placing undue emphasis on the unique needs of an individual at the expense of the common needs of the group? Li- brary users do have unique needs, but on the undergraduate level they have more in common than they have in iso- lation. This is what the "standard works which represent the heritage of civiliza- tion" in the "Standards for College Li- braries" is about. 10 A well-selected collec- tion of books on American history is good anywhere, and not because it hap- pens to serve a particular group of stu- dents or faculty members in Alabama, Alaska, South Dakota, or Minnesota. This paper argues then that the pres- ent system of selection by librarians and faculty members does not produce the quality of collections needed. This is true for several reasons which may be summarized briefly. ( 1) Smaller li- braries cannot rely on Publishers Week- ly, American Book Publishing Record, or Library of Congress proofslips for selection; for if they do, they are buy- ing blind. ( 2) Whether librarians like it or not, under the present system the editors of reviewing journals already C:m- Book Selection I 275 gage in extensive prejudging (selection by inclusion and by omission) both in determining whether a book will be re- viewed and who will review it. ( 3) In order to insure full coverage of current book production, a large number of current general and scholarly journals must be examined . regularly and thor- oughly, and few libraries have the staff time necessary to accomplish such a large task. A practical alternative might be cen- tralized selection on a national basis. The system might work something like this. The Association of College and Research Libraries or ALA's Library Resources and Technical Services Divi- sion would manage the program and would hire subject specialists (twenty individuals with backgrounds in differ- ent disciplines should be able to insure good coverage) who would examine all new books currently being published and who would decide which books were appropriate for the undergraduate lev- el. Depending upon the volume of book production, the twenty specialists would select a total of about 5,000 books per year. The total number would fluctuate with the quality and quantity of publi- cation each year, but 5,000 titles would be a reasonable number for purposes of discussion. This is admittedly a round figur_e, based to some extent on research but also based to some extent on conjecture. A more precise figure could be arrived at as a result of more extensive study. Libraries could buy the package, but they could not make any stipulations about what they would ac- cept or reject. They would take all or nothing. · No exceptions of any kind would be permitted. Such a system could have tremendous side benefits. Why no exceptions? By in- sisting upon a total acceptance of the package, the program could achieve con- siderable economies. Attempting to tai- lor selection to the separate libraries would destroy the program before it 276 I College & Research Libraries • July 1971 had a chance to work, but the package sold to 200 libraries could have tremen- dous economies of scale. One cataloger using Library of Congress copy could supervise the cataloging of 5,000 vol- umes (actually 5,000 for 200 libraries equals 1,000,000 books). Complete card sets could be produced with call num- bers in place. The circulation card and book pocket could also be included. The secret of success would be in the processing of 200 copies of the same book at the same time. Producing 200 sets of cards for one title would permit the use of the best equipment and ob- viously would be far faster and more economical than doing it separately in 200 libraries. One of the major problems in cen- tralized processing is the matter of ex- ceptions. If the processing center allows exceptions, errors are more likely to oc- cur and every member helps to pay for the specialized treatment because excep- tions take time and therefore cost mon- ey. (For a good discussion of the prob- lems of centralized processing see the Fall 1966 issue of Library Resources & Technical Services. ) Furthermore, the simple matters such as spine labeling and producing circulation cards can sometimes be done more economically in the local library than in the central system. However, if this is done en masse, it can be done more economical- ly. It is only when the routine can be done en masse and without a long list of exceptions for each participant that the routines can be done more econom- ically in the central system, for only then does automation provide signi£- c.ant" advantages. A computer, for exam- ple, has an advantage over routine man- ual operations in libraries primarily when the same task must be performed a number of times. If a particular task needs to be done only one, two, or three times, it is likely that the computer will be an expensive luxury. It is interesting to note that in her study of centralized processing centers Vann reported that the buyer-librarian was most likely to be dissatisfied with de- tails of processing rather than with cat- aloging and classification itself-as if the location of the book pocket were the essence of cataloging and usabili- ty .11 Uniform processing for all libraries could ensure a quality product, and it would not cause signi£cant problems for cooperating libraries. It would, of course, mean that all participating li- braries would have to accept the same classification scheme, but this should create no insurmountable difficulties ei- ther. How much would such a plan of cen- tralized selection-acquisitions-processing cost? Broken down by category, it might run something like this: Selection: twenty comple- mentary subject specialists at an average of $15,000 per annum $300,000 Cataloging: one profession- al (should be on the same level as the subject special- ist and his salary could be averaged with that group) $ 15,000 Catalog card sets, including labor and machine costs: $.10 per set for 200 copies of 5,000 titles or a total of 1,000,000 sets $100,000 Processing including all la- bor: $.25 per book for spine labeling, circulation cards, book pockets, match- ing cards with books, etc. $250,000 Administration $ 35,000 Warehouse and equipment: $3,000,000 amortized over twenty years $150,000 TOTAL $850,000 Assuming the .average price of books to be $10.00 per title and an average dis- count of 10 percent per title when pur- chasing. 200 copies of each of the 5,000 l 1 titles, the centralized acquisitions sys- tem would be able to manage all selec- tion, cataloging, and processing for less than the average $1.00 per copy discount ( 200 copies of 5,000 titles equals 1,000,000 books at $10.00 per book equals $10,000,000 and a 10 percent dis- count equals $1,000,000 discount) .12 The total cost of 1,000,000 books, then, would be $9,000,000, and the cost of processing would be $850,000. On a per copy basis this would mean $9.00 per copy and $.85 for processing. Thus the discount would more than cover selection and all processing costs. The library would be able to build a quality collection with the books com- ing to the library ready for the shelves and the cards ready for the catalog at less than the list price of the book. This in spite of the fact that the cost esti- mates above are computed at a rate which is probably higher than they would be in an actual operation. For example, by using Library of Congress copy and offset printing, one worker can easily run 120 cards per min- ute. Using a more conservative average production of only sixty cards per min- ute would mean that one person could produce 3,600 cards per hour ( 600 sets with an average of six cards per set). At $.10 per set this would mean an income of $60.00 per hour. This would allow a generous $.01 for card stock ( $36.00 for 3,600 cards), $8.00 per hour for la- bor, $8.00 per hour for machine rental, and $8.00 per hour for other expenses. Most businesses would be eager .to achieve that kind of return on their in- vestment. Another illustration, the $3,000,000 for the warehouse and equip- ment, is probably high. At a cost of $50.00 per square foot, $1,000,000 would provide 20,000 square feet. Another $1,000,000 for equipment would be quite generous. Thus $1,000,000 wo~ld be left for contingencies. A larger number of subscribers would further reduce the per title processing Book Selection I 277 costs, but even with only 200 subscrib- ers, the smaller libraries could nearly disband their acquisitions, cataloging, and processing centers and invest that money in books. Acquisitions and proc- essing costs in nine Colorado academic libraries averaged $4.09 per book with- out considering institutional overhead. The comparable Colorado Cooperative Book Processing Center cost was $3.10, or $2.96 as calculated in the mathemati- cal model. If two copies of each item could be ordered and processed simul- taneously, the cost of each item would be reduced to $2.27, disregarding institu- tional overhead. 13 A more recent report gives an average cost of $3.10 per book for 1967 and $2.70 during last year's ex- perimental period. 14 .The system de- scribed in this article could perform the same tasks plus the more significant work of selection for $.85 per book. At a cost of $49,250 ( 5,000 titles at $9.00 each plus $.85 each for processing), par- ticipating libraries would have greatly increased their purchasing power by practically eliminating the costs of their cataloging and acquisitions departments. Using the average cost of $4.09 for the Colorado academic libraries and without considering institutional over- head, the processing costs for 5,000 titles would be $20,450. Since the system de- scribed in this article would cost only $4,250 (5,000 titles at $.85 each), each library would save $16,200 in processing costs. Problems with financial records would also virtually disappear for the libraries because bills would come once a month or once a quarter. Billing by the centralized system would also be sim- ple-the same bill would go to everyone. Saving $16,200 in acquisitions and processing costs for 5,000 titles would be no small matter for most libraries. In addition, the library would be as- sured of a higher quality of selection. On any given afternoon, a college's en- tire faculty (including the president) and the library staff could meet in the 278 I College & Research Libraries • July 1971 library, unpack the beautiful new ship- ment of books, and read the books rath- er than the reviews. And eventually, this basic, quality collection could become the minimum acceptable standard for accreditation. Any academic library which cannot purchase 5,000 books per year should not be called a library. Naturally each local library would still need to make provision for unique or additional educational programs and for local materials. However, this would be a comparatively small task. The twenty subject specialists could examine some 30,000 domestic and for- eign titles per year. This would be an average of 1,500 per specialist. Using 200 working days per year as a base, th~s would mean that each specialist would have to look at an average of 7.5 books per day. Assuming a selection of 5,000 titles per year, each specialist would ac- tually approve an average of 250 titles . during the course of one year. Since many decisions for inclusion or exclu- sion would be fairly routine, the spe- cialists should have adequate time to perform their duties. The major duty of the administrator for the centralized system would be to insure complete coverage. It would be his job to make sure that all books which might be relevant to the under- graduate curriculum would get into the system so the subject specialists would have a chance to review them. This would be the critical factor, getting the books into the system for evaluation. Aside from this, tb.e administrator would be responsible for supervising all accounts with publishers and libniries as well as routine tasks such as shipping and receiving. He would also, of course, deal with complaints from librarians. Once the system was operating effective- ly (and he would have to have very hard evidence that it was indeed operating effectively), most complaints could be handled in a fairly routine fashion. If any librarian complained about such matters as the placement of spine labels and call numbers or whether sub- ject entries should be in red rather than in capital letters, the chief administra~ tor for the selection-acquisitions-proc- essing center would write to the com- plaining librarian's ' president (with a copy to the librarian) recommending that the ii)stitution summarily fire the librarian. Why not? REFERENCES 1. }. Periam Danton, Book Selection and Col- lections: A Comparison of German and American University Libraries ( New York: Columbia University Press, 1963), p.71. 2. Bowker Annual of Library and Book Trade Information (New York: R. R. Bowker, 1970)' p.54. 3. Lawrence E. Leonard, Joan M. Maier, and Richard M. Dougherty, Centralized Book Processing: A Feasibility Study Based on Colorado Academic Libraries (Metuchen, (N.J.: Scarecrow Press, 1969), p.171-72. 4. The Book Review Digest has a number of criteria for citing reviews which are ex- plained ·in its "Statement of Policy" in the annual cumulations, but these need not be discussed here. 5. Although the Book Review Digest had not cited it by June of 1970, CHOICE did car- ry a review in March 1970. All the other reviews had appeared in October or No- vember of 1969. 6. Possibly reviews of these works appeared in additional journals without being cited by Book Review Digest or Book Review In- dex, but no attempt was made to do a thor- ough search of a large number of journals to locate additional reviews. . 1. Virgil F. Massman and Kelly Patterson, "A Minimum Budg~t for Current Acquisi- tions," CRL 31:83-88 (March 1970). 8. Leonard, Centralized Book Processing, p.171. . 9. The percentages should not be used as ab- j 1 solute figures since comparisons were drawn for only a small sample. The point to be made is that a substantial percentage of the new books is reviewed by only one journal. 10. "Standards for College Libraries," CRL 20: 276 (July 1959). 11. Sarah K. Vann, "Southeastern Pennsylvania Processing Center Feasibility Study: A Book Selection I 219 Summary," Library Resources & Technical Services 10:472 (Fall 1966). 12. For 1969 · an average price of $9.37 for American books was reported by Publish- ers' Weekly ( 9 Feb. 1970), p.49. 13. Leonard, Centralized Book Processing, p.244. 14. Mountain-Plains Library Quarterly 15:28 (Summer 1970).