College and Research Libraries --- - ~ - - - -~-------------------- H. WILLIAM AXFORD The Econoinics of a Do~nestic Approval Plan A study of the efficiency and effectiveness of a domestic blanket ap- proval plan as compared with traditional acquisitions procedures based upon a unit cost study carried out at four libraries in the state university system of Florida. Comparative costs are measured both in terms of internal systems savings and the level of service provided the academic community. LooKING BACK over the past two dec- ades, one can see that three n1ajor issues seem to have dominated discussions among academic librarians: ( 1) The relative merits of the Library of Con- gress Classification System as compared with the Dewey Decimal; ( 2) the role of the computer in library operations; and ( 3) blanket approval plans as a means for systematic collection growth. All three have generated considerable amounts of impassioned rhetoric at pro- fessional gatherings and a corpus of polemical writing, but only a relatively small amount of research aimed at eval- uating how the Library of Congress Classification System, the computer, or approval plans have advanced or retard- ed the academic library's progress to- ward its stated educational objectives. Nevertheless, the trend toward the Li- brary of Congress Classification System continues to accelerate, the use and abuse of the computer is increasingly evident, and the number of academic Dr. Axford is university lib1:arian at Ari- zona State University, Tempe. This paper was originally delivered at the Third International Seminar on Ap- proval and Gathering Plans in Large and Medium Size Lib1·aries, February 17-19, 1971, West Palm Beach, Florida. 368 / libraries utilizing approval plans grows with each passing year as does the num- ber of firms offering them. This situation suggests at least two possibilities: ( 1) That academic librari- ans may be inclined to rush willy-nilly into anything which carries the magic connotation of being innovative or ex- perimental; or ( 2) that there is some- thing inherently rational about all three of these developments-that they are related to a slowly evolving network for the acquisition, bibliographical control, and dissemination of knowledge on a global scale which overarches the indi- vidual libraries which are its constituent parts. In all probability, both possibili- ties have been operative. But I would prefer to believe that the latter has been by far the most important-that these trends reflect the intuitive genius and pragmatically oriented intellect of the profession probing several promising routes into the future. With respect to approval plans, what is needed at the present moment is a solid body of re- search which will calm some of the con- troversy by moving us from opinion and prejudice into documented facts. In short, we need more than the pro- fession's traditional crutch of self-evi- dent truths or recourse to majority opin- ion to justify what we are doing. We need not, perhaps, go quite so far as a participant in last year's seminar sug- gested and establish an agency similar to the Library Technology Project to study all aspects of this technique of building research collections. Yet surely those of us who are convinced of both the efficiency and effectiveness of ap- proval plans would do well to fol- low Thoreau's advice (slightly para- phrased) , "You have built your castles in the air. Now put foundations under them." Research into the operation of ap- proval plans, though slim in terms of the general interest in the subject, does exist. For instance, for several years the University of Nebraska has gathered very detailed statistics on the number of titles received, the breakdown by LC class, average prices, and discounts. Flor- ida Atlantic University has compiled similar data since 1968. The University of Oklahoma Library has produced a solid vendor pedormance study of ma- jor significance which tested one com- pany's claim "that monographs eligible for coverage under its approval plan would be sent within the same week of publication, and that 80 percent of these would be received before the ti- tle's first appearance in one of the trade bibliographies."1 Finally, there has been at least one dissertation in which ap- proval plans have come under scrutiny. I refer to the work of G. Edward Evans at the University of Illinois in which he compared the use of books received through approval plans and those or- dered individually by librarians and members of the faculty. 2 Much of the published and unpub- lished research on approval plans and the verbal exchanges between propo- nents and opponents share the common attribute of viewing approval plans largely in isolation from the total acqui- sitions and processing effort. This is analogous to designing a powedul new Domestic Approval Plan I 369 automobile engine without facing up to the necessity of also redesigning the entire drive train to achieve the desired level of pedormance. The present study avoids this pitfall since much of the data on approval plans per se has been extracted from a research project intended to derive costs of the technical services divisions of five of the seven li- braries in the State University System of Florida covering operations for fiscal 1968/69. The goal of the study was to measure both efficiency (i.e., the optimal use of human resources) and effective- ness, the level of achievement in terms of established program goals. The methodology of the study was as follows. 1. Each department of the Technical Services Division defined in the clearest possible terms the functions for which it was responsible. 2. A diary study was carried out for each position in each department in or- der to distribute salary / wages and hours worked over the functions performed. 3. The total dollars and minutes spent on each function over a year's op- eration were then divided by the total volumes fully processed by the division. Here it should be noted that the proj- ect was not a true time and motion study, nor was it intended to be. The ob- ject was to determine the average costs per function rather than the determina- tion of a standard time per unit pro- duced. Put another way, we wanted to determine what it was costing to acquire and process a book, including the costs of inefficient supervision, rather than what it should cost. When this proj ect was completed, data were available which made it pos- sible, within reasonable limitations, to isolate cost factors related to books ac- quired through approval plans and those acquired in the traditional man- ner, and to come to some conclusions as to the relative efficiency · of both tech- niques. In order to measure the impact 370 I College & Research Libraries • September 1971 of an approval plan on a library's level of .achievement in terms of its estab- lished program goals, a supplementary study was undertaken which was de- signed to test the effectiveness of an ap- proval plan in expediting current pub- lished scholarship to the academic com- munity. From the unit cost studies of the five participating .libraries, three functions were clearly identified which could be eliminated by utilizing .an approval plan. They were as follows: ( 1 ) Pre- order searching; ( 2) vendor selection; and ( 3) typing purchase orders. Unfortunately, not all of the five li- braries isolated each of the above func- tions in their studies. Consequently, it was necessary in some instances to use the average costs in minutes and dol- lars. For instance, if only four of the five had isolated vendor selection, the average costs for these four would be used for the fifth. Table 1 shows the savings achieved by two of the five libraries in the test group which were on approval plans and the savings which could have been achieved by the other three had they been. Table 2 shows the savings which would have accrued to the state univer- sity system had all five libraries been on approval plans. The range of possible savings be- tween institutions is both interesting and significant. At the lower end of the scale, Library l's figures were 1,073 man- hours and $3,550. At the top, Library 3 could have saved double the number of hours and almost twice .as much money as Library 1 had it been on an approval plan. The greatest variation in costs was in preorder searching. Two factors appar- ently were operative. In the first place, Libraries 2 and 4 which reported the highest costs had more than one F.T.E. professional assigned to this function. The two libraries with the lowest costs had no professional engaged in preor- C')C')t-C')C') t-C0t-COC0 0t-MC0l!) ,....; ,....; c;f c,f ci t-C0lnt-C') C\llnMOO~ c-ic6C')c6~ t-t-C:CC:Ct- C')COOOC') ,....;,....;,....;,....;,....; OOlnlJ)O OOC\100t-t- c-icioo~o5 ... Q) ,.J::l 0 0 z der searching. At Library 1, preorder searching was limited to determining if a given title actually existed. No at- tempt was made to establish the main entry before placing an order with a dealer. The other libraries in the test group followed the traditional proce- dure. The spread in savings actual or po- tential between the five libraries illus- trates an interesting paradox. The more efficient the bibliographic searching and TABLE 2 SAviNGS WHICH CouLD BE AcHIEVED FOR THE STATE UNIVERSITY SYSTEM OF FLORIDA IF ALL FIVE LIBRARIES INCLUDED IN THE STUDY UTILIZED AN APPROVAL PLAN OF 10,000 VOLUMES A YEAR1 Institution Hours Dollars Library 1 1,073 3,550 Library 2 1,733 4,500 Library 3 2,177 6,000 Library 4 2,663 6,800 Library 5 2,533 6,050 TOTAL 10,179 26,900 Translated into positions, the dollar savings would provide approximately five to six clerical positions for the five libraries.2 These figures show that an approval plan, on the average, will save the time of approximately Bf full- time persons. 1 Since over 95 percent of the titles received on an approval plan are single volumes, for the purpose of the study, titles and volumes are considered synon- ymous. 2 The average clerical salary at FAU in 1969/70 was $4,800. acquisition procedures are, the less one will be likely to save by having an ap- proval plan, while the more inefficient they are, the greater will be the savings. The same situation holds true with re- spect to .adjusted discounts; that is, cal- culating the discount on books received on an approval plan to include labor saved. Table 3 shows the results of com- puting the discount on approval plan books on this basis for the five libraries which participated in the study. As can be seen, the adjusted discounts run from a low of just over 11 percent to Domestic Approval Plan I 371 a high of just over 15 percent. Either figure is highly respectable for an acqui- sitions program in access of 10,000 vol- umes covering all areas of study. The evidence derived from the unit cost studies undertaken by the five li- braries of the state university system of Florida clearly support the contention that a blanket approval plan is an effi- cient method of acquiring current do- mestic scholarship. As the data show, a well-managed approval plan can save at the minimum one full-time position, with significantly higher savings possible depending upon variances in internal procedures. The vendor performance study car- ried out at the University of Oklahoma further bolsters the evidence for the ef- ficiency of approval plans. As already noted, the purpose of this study was to test one company's claim that mono- graphs eligible for coverage under the plan would be sent within the week of publication, and that 80 percent of these titles would arrive in the library before their first appearance in a stan- dard trade bibliography. The team which carried out the re- search began by analyzing the fifteen is- sues of Publishers' Weekly from Au- gust 28 to December 2, 1968. These con- tained 8,977 titles. The team concluded that 6,67 4 ( 7 4 percent) fell within the exclusion categories of the library's pro- file, and 2,303 ( 28 percent) within. Of the 2,303 which they felt should have been sent by the dealer, 1, 792 ( 78 per- cent) were located in the library's rec- ords; 509 ( 22 percent) were not located. A subsequent check reduced this latter figure to 466. A list containing these 466 titles was forwarded to the company's regional office to be checked. The district manag- er reported as follows. 1. One hundred and ninety-one of the titles on the list had been con- sidered for inclusion on the ap- proval plan but rejected as being 372 j College & Research Libraries • September 1971 TABLE 3 DEALER DrscoUNTS ON APPROVAL PLAN BooKs CALCULATED To INCLUDE LABOR DoLLARS SAvED THROUGH AN APPROVAL PLAN Average Labor Dollars Saved Additional Discount Total Titles Average Dealer per 10,000 per 10,000 Adjusted Purchased Price Discount per vol. vols. per vol. vols. % Discount Library 1 10,000 $8.93 7.16% $.36 $3,550 $.36 $8.93 4% 11.16% Library 2 10,000 8.93 7.16 Library 3 10,000 8.93 7.16 Library 4 10,000 8.93 7.16 Library 5 10,000 8.93· 7.16 juvenile titles, items of local interest, or nonscholarly. 2. Two hundred and seventy-five had been selected for the approval plan, and of these 111 were judged to fit the University of Oklahoma profile and had been shipped. One hundred and thirty-three were judged as not fitting the library's profile. 3. Thirty titles could not be account- ed for. The dealer's decisions for not send- ing certain titles were based on the ac- quisitions librarian's instructions that the library's profile was to be very strict- ly interpreted. In other words, err in the direction of exclusion in case of doubt. It is possible that the library's inclusion of 111 titles on the search list which had actually been sent under the ap- proval plan may have been due to main entries on the invoices which differed from those in Publishers' Weekly and paperbacks which may have been at the bindery and not located by the library during the study. Subsequent correspondence regarding the results of the study between the company's district manager and the di- rector of the University of Oklahoma Libraries revealed a broad difference of opinion between the director's concept of what the company should be sending and that of . the acquisitions librarian. Whereas the. latter's inclination was to narrow the coverage, th,e former's was to make it as broad as possible. Had the .45 4,500 .45 8.93 5 12.16 .60 6,000 .60 8.93 7 14.16 .68 6,800 .68 8.93 8 15.16' .61 6,050 .61 8.93 7 14.16 approval firm operated under the direc- tor's interpretation of what should come under the library's profile, it is probable that the number of titles re- jected for inclusion would have b een considerably smaller. This situation clearly illustrates a problem which of- ten confronts a dealer. If the library as- sumes a Janus-like stance and speaks out of both mouths at once, the dealer can hardly be criticized if he fails to satisfy either. It also points out the managerial responsibilities inherent in an approval plan. Adjusting the figures to take into ac- count the 111 titles recorded as sent un- der the approval plan but apparently not located in the library at the time of the study, the approval fi1m actually ex- ceeded its claim to deliver 80 percent of the titles which fell within the scope of the University of Oklahoma Library's profile within the week of publication and before their first appearance in a trade bibliography. The .adjusted aver- age early arrival was thirty-one days. After evaluating all the evidence, the research team concluded that the plan was efficiently providing rapid delivery of current domestic publications to the University of Oklahoma Libraries. As work progressed on the unit cost studies, a vendor performance study similar to that done at Oklahoma was undertaken. The methodology decided upon was to take a random sample of the titles received by the Florida Atlan- tic University Library through its ap- prov.al plan during fiscal 1968/ 69 and to check these in the public catalogs of four other university libraries in the state. The libraries in the test group in- cluded a private university which did not utilize an approval plan, a state uni- versity which did, a state university which did not, and a state university which had individual blanket orders with all university presses. In fiscal 1968/69 the Florida Atlantic University Library accepted 10,648 ti- tles through its approval plan. In Jan- uary 1970, six months after the close of the fiscal year, this file was weeded to re- move titles in series, corporate entries, and reprints. An 8 percent sample of the remaining 9,461 titles was then se- lected. Over the cours e of the next five months, the 764 titles obtained by this process were checked in the public cata- logs of the libraries in the test group, the first one in January, two more in F ebruary, and the last the first week of May. It is worth noting at this point that since Library 1 attempted to main- tain a policy of giving original catalog- ing to all approval plan books for which LC copy had not been received Domestic Approval Plan I 373 after ninety days, the bulk of the titles in the sample group had been fully cat- aloged before or by October 15, 1969. The time lags between this date and the dates when the catalogs of the libraries in the test group were checked were as follows. Library 2 Library 3 and Library 4 Library 5 10 weeks 14 weeks 26 weeks The results of the catalog checks were unexpected and not easy to interpret. They are shown in Table 4. Because of the very high percentage of titles not found in the public cata- logs of the test groups of libraries, uni- versity press titles and titles from a se- lected group of individual publishers noted for scholarly publication were studied separately. A slightly different pattern emerged, which again raised as many questions as it answered. For in- stance, at Library 2, which had individu- al blanket orders with all university presses, the percentage of titles not held in the university press group was almost 10 percent higher than for the whole list. Incredible as it may seem, this li- TABLE 4 RESULTS OF CHECKING AN 8 PERCENT SAMPLE OF APPROVAL PLAN BooKs RECEIVED BY FAU DuRING 1968/69 IN THE PuBLIC CATALOGs oF FoUR UNIVERSITY LIBRARIEs IN FLORIDA Level0 2 3 4 5 Total Library 1 Not found 38 1,113 1,550 2,662 4,362 Older ed. Found 0 75 125 125 325 On Order 0 0 0 0 0 Library 2 Not found 38 950 1,162 1,750 3,900 Older ed. Found 0 138 50 88 276 On Order 0 288 Library 3 400 1,061 1,749 Not found 38 1,288 1,688 3,250 6,264 Older ed. Found 12 175 63 188 438 On Order 0 0 0 0 0 Library 4 Not found 38 750 838 1,288 2,913 Older ed. Found 0 63 75 138 276 On Order 0 38 38 50 126 0 Approval finn's assigned level % 56.2 3.4 0 40.8 2.9 18.3 65.6 4.6 0 30.5 2.9 1.3 374 I College & Research Libraries • September 1971 brary did not hold over 50 percent of the university press titles which had been cataloged several months previous- ly at one of the other libraries studied. In order to find out if the libraries not on approval plans were acquiring university press titles by traditional tech- niques, these titles were separated into two groups, those received during the first half of the fiscal year and those re- ceived during the second half. At all in- stitutions, the number of titles not held in the first group was substantially lower than in the second. This finding seemed to indicate clearly that the university press titles which had been received through approval plans were being ac- quired by traditional acquisitions pro- cedures but at a considerably later date . Tables 5, 6, 7, and 8 show the results of this part of the study in detail. In attempting to interpret the data derived from the catalog checks, several points must be borne in mind. First of all, neither Library 2 nor 4 filed "on or- der" infonnation in their public cata- log, nor did they file temporary entries for titles in cataloging backlogs . Con- sequently, it is entirely possible that many of the titles not found in the cat- alog checks were actually owned by these libraries, but not available to the public. Many were probably on order. Although it would have been desirable to obtain this infonnation, not having it does not really .affect the overall find- ings of the study, which revealed a very large number of scholarly titles not available to the academic communities served by these two libraries ten to four- teen weeks after they were available at TABLE 5 NUMBER OF TITLES NOT IN THE PUBLIC CATALOGS OF THE FouR OTHER UNIVERSITY LIBRARIES IN FLORIDA AT THE TIME CHECKED 1 (Number of University Press Titles Received and Cataloged by Library 1-2,137) Institution No. of Titles Percent Library 22 975 45.6 Library 33 1,075 50.3 Library 44 1,325 62.0 Library 55 413 19.3 1 Results of checking university press books received at FA U through blanket approval plan in fiscal 1968 / 69 and fully cataloged by September 15, 1969, in the public catalogs of four other university libraries in Florida. All figures are based on an 8 percent sample of 9,461 titles. All approval plan books for which LC copy was not available after ninety days were given original cataloging at Library 1. 2 Catalog checked in January 1970. 3 Catalog checked in February 1970. Had standing orders with individual university presses. 4 Catalog checked in February 1970. 5 Catalog checked in May 1970. Had blanket ap- proval plan. TABLE 6 NUMBER OF TITLES NOT IN THE PUBLIC CATALOGS OF THE FOUR OTHER INSTITUTIONS AT THE TIME CHECKED 1 (Number of titles received and cataloged from these publishers by Library 1: Praep~r, 275; Wiley, 224; Macmillan, 17 4; Prentice-Hall, 225; McGraw-Hill, 187) Praeger Wiley Macmillan Prentice-Hall McGraw-Hill No. of No. of No. of No. of No. of Titles Percent Titles Percent Titles Percent Titles Percent Titles Percent Library 22 175 63.6 125 55.8 88 50.5 125 55.5 75 40.1 Library 33 113 41.1 75 33.4 75 43.1 125 55.5 13 6.9 Library 4~ 136 49.4 125 55.8 63 36.2 150 66.6 113 60.4 Library 55 13 4.7 0 0 50 28.7 75 33.3 38 20.3 1 Results of checking the above five publishers' titles received at F AU through blanket approval plan during fiscal 1968 / 69 and fully cataloged by September 15, 1969, in the public catalogs of four other university libraries in Florida. All figures based on an 8 percent sample of 9,461 titles. All approval plan books for which LC copy was not available after ninety days were given original cataloging at Library 1. 2 Catalog checked in JRnuary 1970. 3 Catalog checked in February 1970. Had standing orders with individual university presses. ~ Catalog checked in February 1970. 5 Catalog checked in May 1970. Had blanket approval plan. TABLE 7 NUMBER OF TITLES NOT IN THE PUBLIC CATALOGS OF THE FouR OTHER UNIVERSITY LIBRARIES IN FLORIDA AT THE TIME CHECKED 1 (Number of University Press Titles Received and Cataloged by Library 1-1,050) Institution No. of Titles Percent Library 22 387 36.8 Library 33 412 39.2 Library 44 575 54.7 Library 55 150 14.2 1 Results of checking university press bo oks received at Library 1 through blanket approval plan July 1, 1968, through December 31, 1968, and fully cataloged by April 15, 1969, in the public catalogs of four other university libraries in Florida. All figures are based on an 8 percent sample of 9,461 titles. All approval plan books for which LC copy was not available after ninety days were given original cataloging at Library 1. 2 Catalog checked in January 1970. 3 Catalog checked in February 1970. Had standing orders with individual university presses . 4 Catalog checked in February 1970. 5 Catalog checked in May 1970. Had blanket ap- proval plan. Library 1. In the case of university press titles received during the first half of the fiscal year, the time lag was from twenty-four to forty weeks. Libraries 3 and 5 both filed "on or- der" information in the public catalog, and both filed temporary entries for all titles not cataloged. The "on order" in- formation provided some very positive documentation for the effectiveness of an approval plan. At Library 3, the one with individual blanket order plans with all university presses, 18.3 percent ( 1, 7 49 titles) on Library 1' s list were found to be on order. At Library 5, which had an approval plan, the figure dropped to 1.3 percent ( 126 titles). It is interesting but somewhat confusing Domestic Approval Plan I 375 TABLE 8 NuMBER OF TITLES NOT IN THE PuBLIC CATALOGS OF THE FouR OTHER UNIVERSITY LIBRARIES IN FLORIDA AT THE TIME CHECKED 1 (Number of University Press Titles Received and Cataloged by Library 1-1,087) Institution No. of Titles Percent Library 22 587 54.0 Library 33 662 63.0 Library 44 750 68.9 Library 55 262 24.1 1 Results of checking university press books received at Library 1 through blanket approval plan January 1, 1969, through June 30, 1969, and fully cataloged by September 15, 1969, in the public catalogs of four other university libraries in Florida. All figures are based on an 8 percent sample of 9,461 titles. All approval plan books for which LC copy was not available after ninety days were given original cataloging at Library 1. 2 Catalog checked in January 1970. 3 Catalog checked in February 1970. Had standing orders with individual university presses. 4 Catalog checked in February 1970. 5 Catalog checked in May 1970. Had blanket ap- proval plan. to note that both libraries missed exactly the same number of new editions of ti- tles for which they held the previous edition (see Table 4). It seems obvious that the approval plan technique for building research li- braries is here to stay. The study report- ed here clearly demonstrates its efficien- cy and effectiveness. REFERENCES 1. Kathleen Maher, Diana Lane, Martha Schmidt, and Charles Townley, How Good I.s Your Approval Plan, A Vendor Perform- ance Study (University of Oklahoma Li- braries, 1969). 2. Gayle Edward Evans, The Influence of Book Selection Agents upon Book Collection Usage in . Academic Libraries (University of Illinois, Ph.D. diss., 1969).