College and Research Libraries CONSTANCE R. MILLER Scientific Literature as . Hierarchy: Library Instruction and Robert M. Gagne Too often, library use instruction for undergraduates has aimed at teaching high-level concepts, such as how to use an abstract or a citation index, before the essential lower-level skills have been acquired. R. M. Gagne's learning theory points out that emphasis must be placed on the often-ignored prerequi- site of effective library use: an understanding of the systematic way in which scientific (including the social sciences) literature and knowledge is produced, organized, and accessed. This paper establishes the connection between the learning theories of Gagne and systematic patterns in scientific literature, and discusses the implications of this connection for instruction in library use. EARLY ATTEMPTS at library instruction con- centrated on quantity. Learning theories were virtually ignored in presentations that seemed to aim at teaching the entire curricu- lum of a two-year library school course in an hour. In the last decade, however, as instruc- tion in the use of the library to college under- graduates has risen to the status of a disci- pline in its own right, librarians have begun to pay heed to the lessons embodied in the · psychology of learl)ing. For example, self- paced library workbooks designed by Mimi Dudley, Beverly Renford, and others, resem- ble programmed instruction, and a recent article by Kobelski and Reichel reminds in- struction librarians to keep the cognitive the- orists in mind. 1 An interesting speculation, however, is that librarians and learning theorists have been attempting the same thing: to under- stand and to systematize the production of information. Confirmation for this specula- tion can be found by comparing the learning theory developed by Robert M. Gagne (b. Constance R. Miller is instruction librarian, St. John's University/College of St. Benedict, Col- legeville, Minnesota. 1916), a psychologist and instructional de- signer at Florida State University, with the way in which scientific (including the social sciences) information is produced, orga- nized, and accessed. The accuracy with which Gagne's theory of learning reflects the structure of information has implications for library instruction and lends impressive credibility to Gagne's approach to learning. R. M. Gagne emphasizes the influence of learning, rather than growth, on human be- havioral development, 2 and his model of learning is cumulative. This means that progress, or development, results from ac- quiring capabilities that build on each other. Combinations of previously learned items result in new learning. Gagne sees the cumu- lative learning sequence as moving from lower-level capabilities to higher ones (see figure 1), and in his book The Conditions of Learning, he go.es into detail about each type of learning in the sequence, giving examples, conditions for learning, and educational im- plications. 3 Within the sequence, transfer from lower to higher levels occurs because higher-level concepts or rules include ele- ments identical to those at lower levels, plus at least one new element. 4 Learning any- thing new is relatively easy and natural, I 385 386 I College & Research Libraries • September 1982 Problem Solving t Rules t Concepts 1' Multiple Discriminations Chains, mot~ j and verbal~ I Stimulus-response conditions Signalt learning Fig. 1 Gagne's Cumulative Learning Model therefore, if the prerequisite lower-level learning has occurred. It is not only in terms of overall human be- havioral development that Gagne views learning as sequential. He also believes that it is possible to represent, in a valid, ordered se- quence, the instruction of any intellectual skill (i.e., a capability that makes it possible for an individual to execute an entire class of tasks). 5 Gagne calls these valid , ordered se- quences "learning hierarchies": "Any high level skill has one or more immediate descen- dants which are subskills. " 6 Learning hierar- chies do not deal with how something should be taught. They are representations of the in- terrelations of intellectual skills (see figure 2). 7 A learning hierarchy is designed by work- ing backward from a goal or terminal objec- tive with the subordinate skills at each level telling what the learner needs to know in or- der to move on to the superordinate, next higher skill. The lowest skill, or set of skills, in a hierarchy is/are one(s) that all learners can perform. For more advanced learners, instruction can begin at any point in the hier- archy. Task analysis , according to Gagne, is the systematic description of a terminal objec- tive in terms of prerequisite performance ob- jectives. Each skill in a learning hierarchy must be stated in behavioral terms, and in- clude an action verb , conditions for perfor- mance, and a means of measuring whether the performance has been achieved. 8 Design- ing instruction in terms of learning hierar- chies, therefore, has several advantages: it establishes an instructional sequence; it fo- cuses the instruction on the ability or level of each individual learner; and it builds evalua- tion of learner performance and of the in- struction itself into the behaviorally de- scribed system. 9 ' 10 ' 11 ' 12 In spite of the usefulness of Gagne's hierar- Demonstrate "halves" of. a set of objects as 2 equal subsets; "thirds" as 2 equal subsets 't I I Identify 2 apparently Construct subsets equal subsets as which are equal, · "halves" of total; 3 by pairing objects subsets as "thirds" one to each subject of total T T I l Identify the number Deal out objects, Apportion leftover of subsets of a one to each objects, one to group of objects subset each subset as 2 or 3 1' j I I Identify one object I of a set Fig. 2 Example of a Learning Hierarchy chical theory for instructional design, there has been significant opposition to his concept of superordinate-subordinate skill units. The inclusion hypothesis, held by Gagne's oppo- nents, suggests that optimal learning occurs by moving from the understanding of a total task to its composite parts. Ausubel's sub- sumption theory and Norman's web theory maintain that mastery of an overview , or supporting structure, should occur first. 13 Perhaps the major problem with hierarchical learning, however, has been stated by Gagne himself. The validity of learning hierarchies, which is established by proving that learning a subordinate skill implies transfer to learn- ing a superordinate skill, has been difficult to demonstrate. 14 Although relatively recent at- tempts at validation have been somewhat more successful, 15 many studies cast signifi- cant doubt on the whole learning hierarchy concept. 16 This validation problem , how- ever, results primarily from a mistaken ap- proach. An analysis of the organization of scientific literature and of the way in which this organization can be systematically used to obtain information reveals an interesting similarity between library research and Gagne's hierarchical theory of learning. This mirroring of Gagne's theory and the process of accessing knowledge validates the hierar- chical learning model in a way that proving a transfer from lower-level to higher-level ca- pabilities can never do. W. D. Garvey describes science as a "so- cial system of which communication is a sa- lient feature. " 17 This scientific comm unica- tion is farfrom random. It consists, rather, of a complex system 18 of scientific literature, which is organized into networks of interre- lated citations. 19 These networks connect various types of sources on related subjects: there are primary sources, which constitute the unchanging component in a body of sci- entific knowledge; there are secondary Scientific Literature I 387 sources, which are the changing, revising component-the edge of growth; and there are the intermediary or reference sources, which serve as the glue that holds the mass of information together. 20 These latter sources are .the keys to accessing scientific informa- tion. There are two separable but intercon- nected parts to the structure of a body of sci- entific literature, each of which provides a means of conveying what is known. The sub- stantive part provides content, and the bib- liographic part provides bibliographic loca- tion. 2 Figure 3 illustrates the relationships among various sources along a content- bibliographic continuum. 22 To locate primary information on a topic of interest, the content and/or bibliographic sources cannot be used randomly. In the sense that the use of certain sources leads to the use of other sources (e. g. , the location of a key source in an encyclopedia bibliography is a precursor to using a citation index), the sys- tematic way in which scientific literature must be searched to most effectively yield the needed information is hierarchical. Addi- tionally , as Friedes points out, a researcher can "tune in on the scholarly discussion at the level of generality corresponding to his [sic] familiarity with the subject. " 23 The process of developing an understanding of the orga- nization of information is hierarchical , as is Gagne's model of human behavioral devel- opment (see figure 1). To reach the problem- solving stage (i.e., the ability to conduct a complex search for information on a topic), a number of lower-level capabilities must first be achieved. Figure 4 is a representation of the cumulative learning sequence of library research. Equivalent stages from Gagne's cu- mulative learning model are in parentheses opposite the relevant boxes. Gagne applied learning hierarchies not only to behavioral development but also to CONTENT Content and some bibliography Combination of content and bibliographic Bibliography and some content BIBLIOGRAPHIC Dictionaries of definitions only I Longer dictionaries Encyclopedias Journal articles Textbooks Comprehensive Encyclopedias Reviews of Research Handbooks Fig. 3 Abstracts Annotated Bibliographies Continuum of Content and Bibliographic Sources Bibliographies Indexes Citations Indexes Card Catalog 388 I College & Research Libraries • September 1982 (Problem- solving) (Rule) (Concept) (Discrim- ination) (Verbal associatio (Chaining) (Stimulus- response) Conduct a complex search strategy on an assigned topic about which very little is initially known. .0 I Be able to select, and explain the selection, of the type of reference source used at each step in the search process. I Be able to identify a source as "bibliographic" or "content" or both. Define and distinguish between Be able to list and describe the the terms "bibliographic" and main types of ·reference sources "content" as they apply to into which information is reference sources. organized. r Define the term 1 n) l "reference source" r I Describe the purpose of and locate the following parts of a library: circulation, book stacks, Reference area and desk, Periodicals, Media, Card Catalog. I r Know where a library is. Fig. 4 Cumulative Learning Sequence of Library Research the instruction of specific intellectual skills. In the same way, a specific search for infor- mation using content and bibliographic ref- erence sources can be represented as a hierar- chical process. Figure 5 is a learning hierarchy for conducting library research to locate primary information. The hierarchy specifies the systematic order in which sources are used to build an increasingly so- phisticated body of information on a topic. structure of the system of knowledge in which they exist, 24 what applies to the orga- nization of information also applies to the system of knowledge. The ability to hierar- chically gather information on a topic sug- gests the hierarchical accumulation of knowledge in a field. Scientific literature, from the progress of its production, to its or- ganization, to its methods of access, is struc- tured around prerequisite, subordinate steps leading to superordinate steps. The fact that scientists produce and use the literature of their fields cumulatively and hierarchically offers tangible support for Gagne's learning theory. The correspondence between Gagne's the- ory and the organization and accessing of in- formation provides validation for the hierar- chical, cumulative model of learning. To the extent that information sources reflect the Scientific Literature I 389 R. M. Gagne expressed, in explicit psycho- logical terms, a theory of development and learning that accurately reflects a structure implicit in scientific literature. The accuracy of this reflection has implications for library instruction. A student cannot be expected to use an abstract effectively if the student doesn't understand the exact purpose it can serve. Teaching, in detail, the use of Psycho- logical Abstracts to students in a junior-year class will not result in superior research pro- jects. If, however, the students are taught that an abstract, as a type of source, will be useful at the point where their topics are well defined and narrowed, and if they have lo- cated a number of articles, the best of which they need to select, the research projects are likely to represent a creative synthesis of ideas. That is, a student will more effectively use a reference tool of any type by first under- standing the prerequisite concepts: that sci- entific literature is hierarchically organized, TERMINAL OBJECTIVE .... Organize the information on ~ , your topic from the ,~ I primary sources you have located Locate periodical articles through using the periodicals list or Interlibrary Loan. Examine their biblio- graphies for additional sources T Locate books through Use an abstract or an the card catalog or ~ annotated bibliography ~U•e a periodioal index I Interlibrary Loan. located through the guides located through the guides Examine their biblio- to evaluate sources and to find additional articles graphies for additional find additional ones T sources T I' /[ I Use a citation index I _j to locate additional, v 1 more current authors I' . T Use textbook, dictionary L__( \. Use a bibliography or handbook or comprehensive a review of research IL ~ encyclopedia located ~ located through the I' through the guides .1 guides 1' ./1 T p Know quite a bit and Know a little but Know almost nothing and J also know some of don't know any need an overview the major authors on authors names /I\ the subject T T How much do you know about your topic? This will determine the type of source used next. 1' Look in guides to the literature which deal with your topic. T !Pick a topic of interest. START Fig. 5 Learning Hierarchy for Locating Primary Sources on a Specific Topic 390 I College & Research Libraries • September 1982 and that the various elements interrelate. What librarians have to say in support of Gagne, therefore, is that he has described a learning theory that reflects, with great ac- curacy, the way scientists acquire, produce, and access systems of information. In return, what Gagne has to say to librarians is that the key to accessing information lies in utilizing the hierarchical organization of the litera- ture to build a cumulative body of knowl- edge. REFERENCES 1. P. Kobelski and M. Reichel, "Conceptual Frameworks for Bibliographic Instruction," Journal of Academic Librarianship 7:73-77 (1981). 2. R. M. Gagne, "Contributions of Learning to Human Development," Psychological Review 75:177-91 (1968). 3. R. M. Gagne, The Conditions of Learning (2d ed.; New York: Holt, 1970), chapters 4-8. 4. J. W. Cotton and others, "The Identification of Decomposition of Hierarchical Tasks," American Educational Research Journal 14:189-212 (1977). 5. Gagne, The Conditions of Learning, p.237-76. 6. Cotton, "The Identification and Decomposi- tion of Hierarchical Tasks," p.189. 7. Gagne, The Conditions of Learning, p.252. 8. R. M. Gagne, "An Analysis of Instructional 0 bjectives for the Design of Instruction," in R. Glaser, ed., Teaching Machines and Pro- grammed Learning II (Washington, D.C.: National Educational Association, 1965), p.23-26. 9. R. M. Gagne, "Some New Views of Learning and Instruction," Phi Delta Kappan 51:468-72 (1970). 10. M. G. Hackett, Success in the Classroom: An Approach to Instruction (New York: Holt, 1971), p.39-57. 11. H. E. Jones and others, "Hierarchical Learn- ing Paradigm," Journal of Research in Science Teaching 16:489-99 (1979). 12. R. T. White, "Research into Learning Hierar- chies," Review of Educational Research 43:361-75 (1973). 13. Cotton, "The Identification and Decomposi- tion of Hierarchical Tasks," p.198-201. 14. R. M. Gagne, "Learning and Instructional Se- quence," in F. N. Kerlinger, ed., Review of Research in Education (Itasca, Ill.: Peacock, 1973), p.22. 15. R. J. Trembath and R. T. White, "Mastery Achievement of Intellectual Skills," Journal of Experimental Education 47:247-52 (1979). 16. See Cotton, "The Identification and Decom- position of Hierarchical Tasks," p.198; and White, "Research into Learning Hierarchies," p .369. 17. W. D. Garvey, Communication: The Essence of Science (New York: Pergamon, 1979), p.9. 18. Ibid., p.134-35. Garvey has diagramed the · progress of scientific research in a chart, show- ing the stages through which original research must proceed. Interestingly enough, the chart reveals a hierarchical structure made up of prerequisite objectives leading to the terminal objective, the "text, treatise" box. 19. J. MacGregor and R. G. Mcinnis, "Integrat- ing Classroom Instruction and Library Re- search: The Cognitive Functions of Biblio- graphic Network Structures/' Journal of Higher Education 48:17-38 (1977). 20. MacGregor and Mcinnis, "Integrating Class- room Instruction," p.23. 21. This idea is expressed by both T. K. Freides, Literature and Bibliography of the Social Sci- ences (Los Angeles, Calif.: Melville Pub. Co., 1973), p.259-65; and MacGregor and Mcin- nis, "Integrating Classroom Instruction," p.23-25. 22. MacGregor and Mcinnis, "Integrating Class- room Instruction," p .24. 23. Freides, Literature and Bibliography of the Social Sciences, p.260. 24. MacGregor and Mcinnis, and Freides spend considerable time illustrating the truth of this statement. To do a great iob you have to have great tools. Introducing AMI. Which stands for Automated Media Information. You'll probably say it stands for more than that. AMI stands for effortless book- ings, razor-sharp record keeping, and up-to-the- second inventory control. It stands for better inventory turnover, crackerjack customer service and reduced labor costs. To say nothing of reduced labor, period. AMI can do booking searches, by title, subject or borrower. AMI can make reservations far into the future. And it gives you an instant calendar display. AMI can even print confirma- tions, overdue notices, shipping labels, picking lists and usage reports. AMI means the end of old- fashioned man- ual booking. It brings state-of- the-art computer hardware and software to your media center. You get a hard disk com- puter, video display terminal, a printer and the media manage- ment programs you need. All serviced nationwide by NCR. And AMI is expandable. It grows as your library grows. AMI's price? As low as $70 a week. Which means if you've been waiting for the price of automation to come down, your wait is over. Let us give you all the details. Phone RTI now, toll-free at 800/323-7520.* '"""""~" I R~l t RESEARCH TECHNOLOGY 4700 Chase, Lincolnwood, Illinois 60646 *Illinois, Alaska, Hawaii or outside the U.S.A., call312/677-3000.