College and Research Libraries ALLAN D. PRATT Libraries, EconoiDics, and lnfonnation: Recent Trends in lnforination Science Literature Four trends are noted in the information science literature: ( 1) much work in this field is being performed at or for academic libraries; (2) the concept of ''information,, as a process rather than a product is becoming more widely held; (3) both the economic and the concep- tual "value', of information are under serious study; and (4) im- proved measures and models of bibliographic phenomena are being developed which reflect reality reasonably well. ' OVER THE PAST SEVERAL YEARS informa- tion science and library science have been in some sense "competitive" with each other. Though the boundaries of both disciplines are ill-defined, and though each has aspects which are clear- ly distinct from the other, certain ele- ments of each are indeed concerned with similar themes. It is the intent of this paper to demonstrate, through a survey of some of the recent literature of information science, the extent to which the two fields are mutually sup- porting and to suggest that this mutual support may lead to further progress in both. While there is no generally accepted definition of information science, it is possible to adopt a tentative "'definition by enumeration" by referring to the most useful literature survey in the field. A recent (seventh) Annual Re- view of Information Science and Tech- Allan D. Pratt is a member of the faculty of the Graduate Library School, Indiana University, Bloomington, Indiana. nology ( ARIST VII hereafter) pro- vides comprehensive coverage of all those topics which are, in the judgment of the editors, relevant to information science.1 It may therefore be assumed to give a reasonably accurate picture of the interests and concerns of the work- ers in the field; presumably the topics and the relative emphasis among them reflect, at least roughly, their perceived uimportance" in some way. To gain some sense of perspective on the field, it is useful to compare the coverage of different topics in this re- cent volume with that of the first, pub- lished in 1966, ARIST I.2 Broadly speaking, the chapters of both ARIST I and ARIST VII may be divided into four sections: 1. Planning Information Systems and Services 2. Basic Techniques and Tools 3. Applications 4. The Profession In both the first and this recent vol- ume, sections two and three comprise the bulk of the chapters-nine of thir- /33 34 I College & Research Libraries • January 1975 teen chapters, with 78 percent of the to- tal references, in ARIST I; and ten of thirteen chapters, with 85 percent of the references, in ARIST VII. Between the first and seventh volume, however, there have been significant changes in the rel- ative proportions of the sizes of these two sections. The citations in "Tech- niques and Tools" have decreased from 56 percent of the whole in ARIST I to only 25 percent in ARIST VII. On the other hand, the citations in "Applica- tions" have increased from 22 percent to 61 percent of the total. Thus in the course of seven years, the applications literature has grown by a factor of al- most three, while proportionately there is less than half as much literature re- lating to "Techniques and Tools." LmRARIES Much more of the recent applications literature is oriented toward "conven- tional" libraries than was the case ear- lier. Two of the six applications chap- ters in ARIST VII deal with libraries directly: "Library Automation," which has been a fixture since ARIST I, and "Library and Information Networks." Half of the latter chapter's citations ex- plicitly refer to libraries (mostly aca- demic), while the remainder range over such matters as satellite communica- tions, cable TV, the relative merits of different types of equipment, and the like. In addition, a third applications chapter is devoted to "Machine-Read- able Bibliographic Data Bases." Though "data bases" (bibliographic or other- wise) did not appear in the index of ARIST I, there is nearly half a column of such entries in the index of ARIST VII. This interest in bibliographic data bases strongly suggests that libraries are deeply involved in their development and use. The same increased interest in li- braries is evident in two other chapters of ARIST VII. A significant proportion of the citations in both "Costs, Budget- ing and Economics of Information Pro- cessing" and "Library and Information Center Management" refers explicitly to libraries (again, mainly academic). In view of this phenomenon in the literature, it seems safe to conclude that the "conventional" library, far from being displaced by information centers or other new institutions as some pre- dicted, has become the significant "ap- plier" of information science and tech- nology. In the seven years since ARIST I, librarians, or at least those who publish for or about libraries, have grasped the nettle firmly. EcoNOMics A second theme which is becoming ever more prominent in the informa- tion science literature is economics. Threaded through various chapters of ARIST VII are the echoes of the re- mark in the introduction to Part I: "Perhaps the most influential restric- tions on information activities are eco- nomic constraints."3 While ARIST IV and V had chapters on "Library and Information Center Management," as does ARIST VII, the latter is the first to have a chapter devoted to "Costs, Budgeting and Economics of Informa- tion Processing." Though chapters in previous volumes dealt with the evalua- tion of information systems, these were primarily concerned with relevance, re- call, and similar parameters, not cost/ benefit ratios. There is an increasing em- phasis on cost accountability even in those organizations nominally nonprofit, e.g., the requirement that Department of Defense information agencies recov- er half their operating costs from use charges.4 This emphasis is becoming evident also within academic libraries. Buckland has sketched the beginnings of a "the- ory of the library" analogous to the eco- nomists' "theory of the firm," which, if developed, may form some basis for a better understanding of the economics of information.6 Some may feel that he has exercised excessive caution in at- tempting to avoid the assertion that li- braries are a form of business, but his caution is understandable. If he stated flatly that libraries can and should be managed like businesses, the emotional reaction of his readers might well pre- vent their understanding the point he is making. Another obvious indication of this in- creased interest in economics is the for- mation of the Information Industries Association ( IIA) in 1969. This trade association consists of organizations, mostly profit-making in aim if not in fact, most pragmatically concerned with the value of information. One of the more interesting members of IIA is the FIND service in New York, which pro- vides what is essentially telephone refer- ence service to corporate customers. It claims to be able to "cost-account'' ref- erence service, which librarians have traditionally claimed to be either ex- tremely difficult or impossible. 6 The va- lidity of the claim will soon be reflected in FIND's profit and loss statements. Other evidence of the new emphasis on economic matters was the theme of the 1973 ASIS convention: "Informa- tion: Benefits and Costs." This is a good deal more realistic than many previous conference themes, which tended to be somewhat grandiose: "A World of In- formation" ( 1972); "Communication for Decision Makers" ( 1971 ) ; or "The Information Conscious Society" ( 1970). INFORMATION In parallel with the increased empha- sis on pragmatic economics in "real- world" libraries, there has been a nota- ble increase in theoretical attempts to quantify, or at least to define more ac- curately, the "worth" of information. In order to do so, it has first been necessary to overcome the conceptual obstacles presented by the view which Fairthorne termed the "Phlogiston The- Information Science Literature I 35 ory of Information": that "informa- tion is the name of some universal es- sence that can be squeezed out of texts like water from a sponge."7 There is, encouragingly, less tendency than for- merly to speak of "information" as though it were some form of intellectu- al electricity flowing through mental or social circuitry. The distinction between recorded discourse and the process of becoming "informed" is being more widely recognized. That this should happen is under- standable, because it is difficult to make sense of the notion of the c'worth" of information or anything else without considering the question 'Worth to whom, and under what circumstances?" While "the user" has always been an important consideration, at least in a statistical sense, in information systems design, it is only recently that the user as an individual has become an impor- tant element in the direct measurement of "amount of information." One of the early papers recognizing the importance of the individual in this context was Ackoff's in 1958.8 He de- fined three different kinds of messages: those which inform, those which in- struct, and those which motivate a pur- poseful individual. The "amount" of information, instruction, or motivation related to a particular message received by this individual is a function not only of the message but of the individual's prior state. This same point has been re- peatedly made by Fairthome in various contexts. It was also made, in a tentative and informal way, by Pratt in 1967.9 Mitroff, Williams, and Rathswohl have constructed a model of a ''Dialec- tical Inquiring System,"10 the mathe- matical formulation of which is based on Ackoff's 1958 paper. They indicate divergencies in their viewpoint from that given in Yovits' 1969 paper in which he proposed that "information is data of use in decision making," and sketched out a model based on that defi- 36 I College & Research Libraries • January 1975 nition. 11 However, the more recent work by Whittemore and Yovits suggests that the two conceptual structures are rather close, at least in some major respects. 12 There are strong similarities between Mitroff's "purposeful individual" who is moved from state sl to state s2 and the D M (decision maker) of Whitte- more and Yovits. The latter authors sug- gest six different varieties of uncertain- ty, none of which correspond to the classic "uncertainty" of the Shannon/ Weaver model. What Whittemore and Yovits call "executional uncertainty" is broadly the same as Ackoff' s "efficiency" variable incorporated into Mitroff' s model. The differences between these two conceptions seem to lie, at least in part, in the differing underlying aims. Mitroff speaks of "inquiry" with its connotation of more or less detached and philosoph- ical ends, while Yovits' concept of a DM has a much more aggressive implication. It is rather as if the former is trying to model Bertrand Russell while the latter is trying to model J.P. Morgan. Other indications of this more direct involvement of the individual in the "information" process can be found in recent work. Cooper proposes a concep- tually simple, though methodologically difficult, measure based on the point that "there can be in principle no better evaluation . . . of a retrieval system's benefits than a subjective evaluation by its users, provided that such an evalua- tion is made with all due care and is measured judiciously in terms of an ap- propriate utile."13 Some means of mak- ing his subjective evaluation method op- erationally viable are described in a more recent paper.l4 Simon, in a quite different context, brings up the same general theme.15 His point is that the critical resource is not "information" (which Yovits would call "data") but attention-that the decision maker (or inquirer or user) cannot af- ford to devote enough of his limited time to "attend to" all the data avail- able. Thus, effort needs to be placed not on data gathering, but on organizing it. "Memories [including libraries, etc.] as components of information-processing systems, need to be viewed as stores of potential information, which, if in- dexed effectively, can become available at a reasonable cost whenever it is need- ed as input to a decision-making pro- cess."16 (Emphasis in original.) MEASUREMENT Some recent applied mathematical work is progressing along parallel, or at least related, lines. There is, in addition to the mathematics associated with the cost-accounting problems noted earlier, a growing tendency to concentrate on the mathematical description of phe- nomena which are in fact quantifiable, a change from efforts in earlier years which purported to measure "relevance" to three decimals. One seems on firmer ground when measuring citations, vol- umes, journals, circulations, and the like than when measuring relevance. In particular, there seems to be a growing belief that many bibliometric phenomena and their interrelationships may be described by some forms of ex- ponential or logarithmic functions. The most general statement along these lines is Price's remark that . . . it turns out roughly that the true measure of value is not measured by the crude magnitude, but by its log- arithm. A collection of related books or journals or papers gains the same degree of utility in going from 100 items to a thousand as in going from 1000 to 10,000 or from a million to ten million. 17 This characteristic has been implicit for some time in the various studies of the distribution of journal articles in certain fields, under the general name of the Bradford-Zip£ distribution. Wil- liams found that the demand for peri- odical titles was well described by the lognormal distribution (in which not the value of the variable under consid- eration, but its logarithm to some con- venient base-usually e or 10-is nor- mally distributed) .18 Palmour, Wil- liams' co-worker, has stated that the same function predicted demand for documents at NTIS as well. 19 Resnikoff and Dolby have found that the five-year circulation pattern of Har- vard's Widener Library approximates a lognormal distribution fairly wel1.2° (Incidentally, it was found that only 6 percent of the collection had circulated at least once during that time.) Res- nikoff and Dolby have also used the Whitworth distribution, the discrete analog of the continuous logarithmic function, to describe the distribution of books over subject classes.21 Relation- ships between the Log-Log (Zip£) dis- tribution, the Whitworth, and the log- linear have been pointed out by Griffith and Krevitt.22 Naranan has generalized some of these distributions in terms of power laws.23 These forms of distribution have been found useful, not only in studies of journals and citations but in circula- tion studies as well. In addition to the Resnikoff and Dolby work noted ear- lier,24 Trueswell found similar distribu- tions in his study of libraries at North- western. 25 He was able to make reason- ably accurate predictions of collection use from his models, and found that something in excess of 50 percent of the collection could be removed with little impact on the user satisfaction lev- el. It appears that these power-function distributions underlie a large range of phenomena related to bibliographic and information science phenomena. They are of interest not only because they may provide some empirical support for the theories mentioned above but be- cause they may have direct applications in quantifying some aspects of librari- anship. Information Science Literature I 31 There has been no recent explication of these distributions published for the benefit of those who are more literate than "numerate." Fairthorne,s historico- bibliographic essay of 1969 is the only paper which summarizes work up to that time in reasonably nonmathemati- cal terms. 26 There is need of a similar effort covering the last five years because if the applied mathematics is to be in fact applied, it must be understandable to those who would apply it. SUMMARY This examination of recent informa- tion science literature suggests the emer- gence of four trends which may be in- dicators of new directions: 1. Much of the research and develop- ment work in information science is being conducted by, for, or un- der the direction of ''convention- al" (mainly academic) libraries. 2. There is increasing awareness that "information" is not a property of documents but a complex relation- ship between (at least) message and recipient. 3. There is serious consideration be- ing given, on both a theoretical and an applied level, to the ques- tion of the "worth" of informa- tion. 4. There is some progress being made in bibliometrics-the development of means whereby one can measure and predict library-related phe- nomena with models which reflect reality reasonably well. These trends, if continued, may raise information science from the low level to which Saracevic felt it had fallen by 1971. In five years of work, he asserted, "The basic literature information prob- lems were not solved, or even dented, especially in comparison to the efforts and money invested."27 There is room for cautious hope that some small dents may soon appear. 38 I College & Research Libraries • January 1975 REFERENCES 1. Carlos A. Cuadra, ed., Annual Review of Information Science and Technology [Here- after referred to as ARIST] Volume 7, 1972 (Washington, D.C.: American Socie- ty for Information Science, 1972). 2. Carlos A. Cuadra, ed., ARIST, Volume 1, 1966 (New York: Wiley, 1966). 3. Cuadra, ARIST, Volume 7, 1972, p.4. 4. Ibid., p.39. 5. Michael Buckland, "Toward an Economic Theory of the Library," in Robert S. Tay- lor, ed., Symposium on the Economics of Information Dissemination (Syracuse, N.Y.: Syracuse Univ. Pr., in press). 6. Paul Doebler, "Seek and Ye Shall FIND," Publishers' Weekly 202:79-80 (Oct. 16, 1972). 7. Robert A. Fairthorne, "'Use' and 'Mention' in the Information Sciences," in Laurence B. Heilprin, Barbara E. Markuson, and Frederick Goodman, eds., Proceedings of the Symposium on Education for Informa- tion Science, Warrenton, Virginia, Sep- tember 7-10, 1965 (Washington, D.C.: Spartan Books, 1955), p.10. 8. R. L. Ackoff, "Towards a Behavioral The- ory of Communication," Management Sci- ence 4:218--34 ( 1958). 9. Allan D. Pratt, "An Essay on the Nature of Information," Proceedings. American Documentation Institute. Annual Meeting, Volume 4. (Washington, D.C.: Thompson Book Company, 1967), p.3-7. 10. Ian Mitroff, James Williams, and Eugene Rathswohl, "Dialectical Inquiring Systems," Journal of the American Society for In- formation Science 23:374-78 (Nov.-Dec. 1972). 11. M. C. Yovits, "Information Science: To- ward the Development of a True Scien- tific Discipline," American Documentation 20:374-76 (Oct. 1969). 12. Bruce J. Whittemore and M. C. Yovits, "A Generalized Conceptual Development for the Analysis and Flow of Information," I ournal of the American Society for I nfor- mation Science 24:221-31 (May-June 1973). 13. William S. Cooper, "On Selecting a Mea- sure of Retrieval System Effectiveness," Journal of the American Society for In- formation Science 24: 100 (March-April 1973). 14. William S. Cooper, "On Selecting a Mea- sure of Retrieval Effectiveness. Pt II," journal of the American Society for Infor- mation Science 24:413-24 (Nov.-Dec. 1973). 15. Herbert Simon, "Applying Information Technology to Organization Design," Pub- lic Administration Review 33:269- 70 (May- June 1973). 16. Ibid., p.272. 17. Derek J. De Solla Price, "Some Remarks on Elitism in Information and the Invisible College Phenomenon in Science," Journal of the American Society for Information Science 22:74 (March-April1972). 18. Gordon Williams et al., Library Cost Mod- els: Owning Versus Borrowing Serial Pub- lications (Chicago: Center for Research Li- braries, 1968). ( PB 183 304). 19. Vernon Palmour, Personal communication, June 21, 1972. 20. H. L. Resnikoff and James L. Dolby, ACCESS: A Study of Information Storage and Retrieval with Emphasis on Library Information Systems (Los Altos, Calif.: R & D Consultants Company, 1971). (ED 050 773). 21. H. L. Resnikoff and James L. Dolby, "On Archival Access," in Jeanne B. North, ed., Proceedings of the American Society for Information Science, Volume 7: The In- formation Conscious Society (Washington, D.C.: American Society for Information Science, 1970), p.255-64. 22. Belver C. Griffith and Beth Krevitt, "A Comparison of Several Zip£-Type Distribu- tions in Their Goodness of Fit to Language Data," Journal of the American Society for Information Science 23:220- 21 (May- June 1972). 23. S. Naranan, "Power Law Relations in Sci- ence Bibliography-A Self-Consistent In- terpretation," Journal of Documentation 27:83-97 ( 1971). 24. Resnikoff and Dolby, ACCESS. 25. Richard W. Trues well, "Determining the Optimal Number of Volumes for a Li- brary's Core Collection," Libri 16:49-60 ( 1966). 26. Robert A. Fairthorne, "Empirical Hyper- bolic Distributions ( Bradford-Zipf-Mandel- brot) for Bibliometric Description and Pre- diction," Journal of Documentation 25: 319-43 ( 1969). 27. Tefko Saracevic, "Five Years, Five Vol- umes and 2345 Pages of the Annual Re- view of Information Science and Technol- ogy," Information Storage and Retrieval 7:138 (Sept. 1971).