College and Research Libraries Searching Techniques in the Literature of the Sciences THERE ARE MANY points to consider in making an effective science literature search. Initially, it is assumed that one must attempt to locate all available in- formation relevant to a subject at the level required by the requester. ·with patents, all available information is re- quired in order to establish priority in the recording of ideas and inventions. With the volume of literature now grow- ing at an accelerating and unprece- dented rate, retrieval is made more dif- ficult by the sheer volume of information available. An effort to retrieve specific information concerning a specific sub- ject or an effort to retrieve general in- formation concerning a general subject must be carefully planned. Efficiency in literature searching can be gained only by experience, but certain principles which may be modified to suit the con- ditions of almost any particular situation can be set forth. In order to perform a comprehensive examination of literature, the following steps should be taken: l. Determine the subject scope 2. Determine the temporal scope 3. Determine the depth 4. Determine the literature to be used, e.g. ~ abstract journals, handbooks, books, etc. 5. Familiarize oneself with the indexes to be used 6. Determine the subject headings in- volved 7. Scan the subject indexes 8. Check the abstracts cited In the in- dexes for pertinency 9. Prepare a list of authors of selected references MAY 1963 BY HUGH E. VORESS Mr. Voress is Chief of the Techn"ical Lit- erature Section, Division of Technical In- formation Extension, United States Atomic Energy Commission. This article was pre- pared from a talk given by Mr. Voress at the Conference on the Literature of Nuclear Science: Its Management and Use, held at Oak Ridge~ Tennessee , September 11-13, 1962. 10. Scan the author indexes for addi- tional references 11. Prepare a list of corporate a~thors (issuing agencies) 12. Scan the corporate author indexes for additional references 13. Arrange all the references in the or- der desired 14. Edit the citations into the same for- mat 15. Compose the finished copy. ExAcT NEEDS OF THE REQUESTER The searcher must first determine the scope of the subject. In other words, he must determine "what the requester wants." Although a written statement should be obtained of his needs, it is frequently necessary to have several dis- cussions with the requester in order that these needs might be understood clear- ly. Requesters may not know just what type of information will be helpful or available, or even what type of informa- tion they want. Such a discussion may help to formulate better the exact re- quirements. The requester should be asked if he has done any investigating himself in order to avoid duplication of effort. If he does not understand how to 209 examine the reference tools he has al- ready used, this work may have to be done again by a trained searcher. The parameters or limits of the re- quester's needs often go beyond subject alone. It may be that only information developed after a certain date is sought. Or, he may want only that developed by a certain person, company, or in certain countries. Adding qualifications of this type will greatly reduce the amount of time and effort required for retrieval. ScoPE oF THE REQUESTER's NEEDS The definition of the scope of the re- quest should be explored from a sub- ject standpoint so that all possible in- formation related to the request will be considered. An experienced searcher ex- amines textbooks, encyclopedias, and dic- tionaries to find related subjects of prob- able interest. No one, regardless of his subject specialties, can know all of these relationships. Discussion with a scientist or specialist in the field, other than the requester, often serves to determine whether any fields have been overlooked. SouRCES AND R EFERENCE TOOLS TO BE USED In making the search itself, the first step is to examine various guides to sci- entific literature in the reference section of the library for selection of the proper reference tools. Those offering the most promise should be chosen before work is begun. Use of such a publication as Sci- ence R efer ence Sources by Frances Briggs Jenkins will assist in making this deter- mination. HEADINGS TO BE ExAMINED It is necessary in the case of abstract journal indexes to understand the par- ticular method of indexing used. This is usually explained in the introduction to the index. The United States Atomic Energy Commission issues a publication entitled "Subject Headings Used by the Atomic Energy Commission," TID-5001, which provides such an explanation. Constituting the subject heading au- thority for Nuclear Science A bstractsJ TID-500 I lists permissible headings that can be used by the indexer, and cross- refers to other related information. This publication has been distributed to all contractor and depository libraries. Pertinent subject headings vary from one abstract journal to another, and it has been found helpful to list the sub- ject headings selected for each journal. Revision of these lists becomes necessary as the work progresses and relationships are discovered that were not readily ap- parent at the outset. CoMPREHENSIVENEss AND TIME PERIOD A fundamental decision that must be made is whether the investigation needs to be brief or comprehensive. Occasion- ally, a preliminary determination is made of the amount of literature that will have to be covered. This will permit the searcher to advise the requester of the probable time involved. In general, comprehensive subjects usually require months of work, and a requester should be advised of the probable cost in both time and effort. Just how comprehensive the review should be is only one facet of the problem; another concerns the span of time in months and years to be covered in the literature. A third con- sideration is the body of literature that should be covered, i. e .J patents, periodi- cals, handbooks, books, monographs, bib- liographies, reviews, encyclopedias, etc. Review articles, books, monographs, and bibliographies frequently summarize old- er information and thereby may reduce the necessity of reviewing some of the older literature. It is helpful to arrange these sources of information in order of importance; then, should expediency be 210 COLLEGE AND RESEARCH LIBRARIES ~. a factor, the most important source can be checked first . Usually, the most re- cent sources are explored first. EVALUATION OF REFERENCES AND ABSTRACTS After an index has been searched, the cited abstracts must be read to determine whether they are pertinent. This is nec- essary because all people do not think alike, and methods of conveying infor- mation through indexes are often not consistent. Indexes cannot fully express the content of titles and abstracts. Should the information uncovered be of a particular nature, e.g. ) patents, scan- ning of publications specializing in pat- ent literature may turn up additional references. AuTHORs AND CoRPORATE AuTHORS Another aid is to list the authors of the selected references. Examination of the author index in an abstract journal may result in the location of additional references that were not indexed. Some- times an examination of the corporate authors or issuing agencies can also be fruitful. Work being performed at an installation may be subject-indexed by several different people over a period of time . This can result in variations in style, as a subject considered worthy of indexing by one person may not be con- sidered so by another. A search of the corporate au thor index may give addi- . tional help in alleviating this problem. STYLING OF THE REFERENCES Editing of all references found into a single style is the next step. This is ad- visable because there are many indivi- dual styles used by the various abstract- ing services. Almost all abstract journals vary as to journal abbreviations in the citation. However, the same abbrevia- tions should be used for references from a given journal source. In addition, some MAY 1963 journals may list references by author, while others may cite by title. TERMINATION OF SEARCH During styling and consolidation, if evaluation indicates that the informa- tion desired has been located, this phase may be terminated. However, location of only scanty amounts of information may indicate that a greater time span may be necessary or that a different body of literature should be surveyed. As men- tioned before, the most important ab- stracting journals pertinent to the sub- ject should be examined first, starting with the most recent issues. As one ex- amines the older issues of an abstract journal, it is important to remember that there may be changes in nomen- clature and terminology which will ob- viously affect subject headings. PRESENTATION OF REFERENCES A final point to consider is the man- ner in which the information will be presented to the requester. Within the USAEC Division of Technical Informa- tion Extension, any literature search which has taken appreciable time or has resulted in a significant list of refer- ences is published in order to provide this material to others with similar needs. These publications are distributed both to the Atomic Energy Commission con- tractors and to the depository library col- lections. They are also cited in the TID- 3700 series of Informal Listing of Bibli- ographies of Atomic Energy Literature, which covers on a current basis bibli- ographies in preparation and completed. Bibliographies of Interest to the Atomic Energy Program) TID-3043 (Rev. 2), is a cumulation covering all bibliographies and literature searches through Novem- ber 1961 and is supplemented on an an- nual basis. These documents are in the depository collections and should be kept in the reference section. The latest 211 complete list of AEC depository libraries appears in current issues of Nuclear Sci- ence Abstracts. question, the sources and reference tools most likely to be productive, subject headings to be examined, and the time period to be covered. Necessity for eval- uation of the references and abstracts located, additional searching of author indexes, and final selection, styling, and presentation of references are discussed. Criteria for terminating searches are shown to vary according to the condi- tions of the individual search. SUMMARY The steps employed in searching the scientific and technical literature at the Division of Technical Information Ex- tension, AEC, Oak Ridge, Tennessee, are described. Determinations must be made concerning the exact needs of the re- quester, the scope of the su bject area in Argonne National Laboratory Tour ACRL Subject Specialist Section-RSD Information Retrieval Com- mittee-Interdivisional Committee on Documentation are sponsor- ing a tour of the library and information services of the Argonne National Laboratory on Friday afternoon, July 19, during the Chi- cago conference of ALA. . Buses will leave the Hilton Hotel at 1:30 P.M. and will return passengers to the hotel at 6:00 P.M. Cost for transportation will be $1.50 per person. The tour will include the library and information resources, plus some of the facilities such as a nuclear reactor of the Argonne N a- tiona! Laboratory, and the Applied Math division where computer use in library procedures will be demonstrated. Tickets for the tour will be available at the conference Tours desk. • • Grant for Columbia Operations Analysis A GRANT of $7,200 has been made by the council on Library Re- sources to Columbia University to assist the university libraries in an operations analysis, looking toward improvement of administra- tion. The investigation will have the assistance of some nonlibrary university staff and facilities, including computers if needed. The investigation will seek to identify the major activities of research library operation, the kinds and levels of skills of competency in- volved in each activity, mental versus manual processes-all this in relation to other activities and to the operation as a whole-and what may be gained or lost if changes are made in these activities. • • •• 212 COLLEGE AND RESEARCH LIBRARIES '