SUMMARY OF DISCUSSION
DR. PHILIP M.MORSE, Panel Chairman, opened the discussion by stating that this first panel session concerning the informational needs of scientists in a way would set the stage for discussions of methods of meeting these needs later in the week. The question of the use of scientific information as a tool for research would be the central theme in this discussion. The quantity of scientific information is immense and is expanding exponentially. Its transmission from source to user is multi-channeled and very uneven. It cannot be concentrated into one or even a few types of channels. The papers imply that the greater the distance between fields of specialization, the less the efficiency of communication channels. Interdisciplinary communication is an extremely important problem to solve.
Some of the papers seem to indicate that scientists prefer face-to-face communication, which has a number of advantages. Would it be possible in the more mechanical methods of communication to incorporate a method of direct questioning and to select that part of the answer that is of immediate interest, as can be done in face-to-face communication?
Much scientific information is perishable goods, not treasure to be stored indefinitely. Material should not be left in the general research library too long.
Dr. Morse proceeded to discuss the paper by Halbert and Ackoff which reports on an attempt to see whether the field of operations research can help in this immensely complicated process of transfer of information. It is a brave effort that needs to be continued, but there is a long way to go. Whether the measures and criteria used are the best will have to be decided as time goes on. A number of the results of the study are of considerable interest, particularly the finding that chemists spend most of their working or productive time communicating, a finding that strengthens the belief that scientific communication is an important field. Whether the results are typical, even for chemists, is perhaps a question the reader can decide for himself. One can inquire whether operations research really has much chance of helping in this field, that is, whether the “operations” and interactions between the “system” and the people using it are sufficiently amenable to statistical analysis that operations research has much chance of being successful.
Dr. John W.Tukey pointed out that in the operations research approach there seems to be only an attempt to see what scientists do and not to ask what they think about things—and this in a process that is full of people.
Dr. Morse replied that he thought Dr. Tukey would agree that most of the so-called market research is really beside the point. One must take people’s opinions about things with a grain of salt. Until now there has not been enough effort to find out what people do and too much of asking them what they think and want. Actually we should not have either type of study without the other.
Dr. Russell L.Ackoff commented on the question as to whether the results of the studies reported in the Conference papers are generalizable. In the Case Institute study, the population was well defined and the sample was selected in a precise way so that it would be possible to make some generalizations about the population of 55,000 chemists. He also stated that numerous examples can be given to illustrate that observational techniques obtain much different and more reliable data than questions asked of individuals about their behavior. It does seem to be important to get methods of observation that are subject to measures of reliability. As for the question of interdisciplinary communication, a profile of the chemist can be drawn from the results of the Case study that indicates that as he becomes a minority in a group he resorts more to written communication.
Dr. Mortimer Taube pointed out that the paper on the Case study stated that within the limits of the study it was not possible to develop a measure of effectiveness for scientific research. Instead the criterion of the availability of more time for research was used. If time is the only consideration, the best way to give scientists more time for research would be to abolish publications and information services so that they could spend all their time in the laboratory, but would science then be really productive? He suggested that there is confusion in the paper between the notions of productive science and scientific productivity.
Mr. Michael H.Halbert replied that the Case paper was not meant to convey the impression that scientific productivity is time spent in the laboratory. Scientific productivity is made up of the set of activities categorized in the study, including scientific communication, equipment set-up and use, data treatment, thinking and planning alone, etc. The scientist has only 100 per cent of his time to allocate. If one can find those factors that affect the way a scientist allocates his time, it might be possible to show that certain changes would make more time available for the scientist to use in whatever types of scientific activities he thinks best. It appears to be the consensus of chemists that it is most productive to allocate the largest proportion of their time to communication.
In speaking briefly of the Glass-Norwood study, Dr. Morse commented that many persons have a rather deep-seated suspicion of questionnaires concerning
preferences. The Glass-Norwood study avoided this difficulty by asking scientists only about actions in specific cases. The results are of interest but only fragmentary.
Dr. Bentley Glass commented that the thing that came out of their little study that is most disconcerting and at the same time illuminating is the fact that the great majority of scientists questioned indicated that general conversation was the most frequent source of information about work that was essential to the development of their own ideas. He talked with a good many other scientists after the study was done and found that none of them disagreed with this feeling about sources of information. This finding implies that our existing sources of information, such as abstracting and indexing services, are not fully utilized, and it indicates a great degree of provincialism among American scientists. Dr. Glass said he cannot believe that the obtaining of essential information through casual conversation is a good means of obtaining knowledge of all that is going on in the world of science.
Dr. Chauncey D.Leake briefly discussed the Menzel paper pointing out that it clearly states the objectives of the interview survey at Columbia University. All channels of information exchange and all functions which scientific communication facilities are called upon to perform were considered. One point of interest is the significance of the unplanned mechanism, when a scientist while searching for a particular item stumbles upon something else useful. The question raised is “Are these cases merely individual accidents or do they occur with aggregate regularity?” Another question, that of motivation, which was mentioned by Dr. Menzel, merits further consideration. An important question to ask is “for what?” We so frequently take our objectives for granted that perhaps it is wise to take a critical look occasionally at our assumptions. Scientific information has manifold values, in logics, aesthetics, and ethics. Varying motivational factors are involved in producing, storing, and retrieving scientific information, but always the question is “for what?” If we can agree on some of the goals and objectives, then certain consequences are inevitable. These comments lead to a practical suggestion. Our modern concepts of science and democracy are derived from John Locke. One of his primary principles was that if persons in the community have access to the same body of reliable information, they will make wise decisions on policy matters. This principle was extended by Benjamin Franklin to the obligation of society to provide a free public library and by Jefferson to the provision of free public schools. Is it not time to extend the principle further to the obligation of society to provide for free distribution of scientific information at least to public and university libraries? Actually such distribution may become essential if the free flow of scientific information is to continue.
Dr. Y.S.Touloukian added the question “for whom” to Dr. Leake’s question “for what.” It is essential to distinguish between the generators and users of information, who are distinct in their needs and characteristics.
Dr. Leake agreed that Dr. Touloukian was wise to raise the question, “For whom?” In spite of all that is said about the importance of making contact with those in other fields, scientists are still talking largely to each other. Those who interpret scientific developments for the benefit of people generally may be performing as important a scientific function as scientists themselves. Dr. Leake went on to suggest that an attempt be made to apply measurement techniques to motivations.
Dr. Leonard Carmichael commented that Professor J.D.Bernal in his paper had presented some of the results of his long and thoughtful studies of almost the whole field covered by this Conference. It is written from the point of view of the user and emphasizes the importance of a natural historical approach to the problem. He makes some very concrete proposals as to how the processing of scientific information may be carried out in an effective manner. He points out that knowledge of the requirements of users and of the uses to which they wish to put the information should be the ultimate determining factor in designing methods of storage and retrieval.
Dr. Carmichael went on to say that the point of view of the physiological psychologist, or human engineer, should not be forgotten in considering the problem of scientific communication. For example, the rate of silent reading of standard prose by university students of equally good scholarship has been found to vary greatly (from 2.5 to 9.8 words per second). There are many individual differences in all aspects of information reception and use by human beings, and also many differences in learning and retention abilities. Neurones as well as electronic devices are concerned in the input, storage, search, scanning, synthesis, recall or retrieval, and recognition of scientific information. Human perception and motivation are both better understood today than they were ten years ago, and this knowledge is relevant to the whole problem of the nature and use of scientific information.
A book entitled Reading and Visual Fatigue, written some years ago by Dr. Carmichael and Dr. Walter F.Dearborn, was based on prolonged experimentation. Among the findings of this study was the fact that reading of microfilm does not produce more eye fatigue than book reading; nor is the speed of reading of microfilm necessarily different. This finding among others has relevance to the problems discussed by this Conference. Consideration of sensory processes, central nervous system activities, and human motor processes should not be neglected for they are basic in information production and use.
Professor Bernal emphasized that it is important to make a clear distinction
between the people who make inquiries and the kinds of inquiries they make. Information about both is basic to the design of information systems. History suggests that the different types of inquirers and inquiries should be analyzed by subject. The habits of different kinds of scientists are dictated by the nature of their work. Separate studies of the part scientific communication plays in every scientific discipline are needed.
Professor Bernal then said there appears to be a very large disparity between the edifice of scientific communication and the practices of research scientists. He had thought that the edifice was built primarily for practical scientists, but the studies indicate that they do not read very much anyway. Possibly an enormous edifice is being built for no particular use at all. Scientific information is pouring out at a rate that is hyperexponential, and reading time, however well it is spent, cannot be increased by a factor of more than two. Something quite drastic has to be done about this situation. There is just too much relevant information. Scientists can perhaps continue what they have always done—work with the aid of hunches, luck, conversations, etc. Something a little more positive, however, ought to be introduced into our scientific information services. More effort might be made to direct scientific information to people who can make use of it. One way would be to make more use of the intelligent information officer who knows the people he serves, as long as he doesn’t overdo it and direct their attention to so many things that they are worse off than they were before. Another method would be to go right back to the seventeenth century and use personal contacts more extensively.
Dr. W.F.Libby commented that the Spirit-Kofnovec paper describes a routing system for information by information officers that is fairly standard in many organizations, including the U.S. Atomic Energy Commission, which distributes its reports by categories.
Dr. Libby went on to say that the analysis of reference questions by Herner and Herner contains several very interesting implications. Any arrangement of knowledge must cover to some degree all of the disciplines. The main call on information services, as Professor Bernal has put it, is not in respect of pieces of information that are central to the work of the inquirer, but rather ancillary to it. The worker has means of finding information in his special field of study. The Herners were wise in cautioning readers of their paper that reference questions do not encompass all the requirements of research workers for information. Their analysis of the number of concepts involved in reference questions and the logical relationship between them suggests that the development of information retrieval systems and machines may have been delayed because the designers are trying to anticipate every possible complexity that might be encountered. It may be that relatively simple systems and devices will suffice.
Studying the information needs of scientists is very confusing. Dr. Libby emphasized that such needs vary with the subject, the type of research (pure or applied), the type of user (research worker, historian or writer), the purpose for which the information is sought, the availability of information services, the amount of material on hand, language capabilities and general ability. All these factors make for a confusing situation in studying the information needs of scientists. It appears that all forms of information services are used.
It is apparent from most of the studies that research workers rely greatly on oral communication. Students are required to learn how to use the literature, but it is apparent that as a scientist becomes older and better established he relies more and more on oral communication. It is well known, however, that some of the most productive years of scientists are the younger years. Reliance on oral communication may not be the best technique. On the other hand, the researcher is often seeking information at the forefront of a developing field that is not in the library, and he must get it by letter or conversation or at meetings.
Until now, Dr. Libby said, the abstract journal has not been an ideal medium for searching the current literature because indexes could not be prepared on a current basis. Therefore beginning in January of next year Nuclear Science Abstracts, which is published by the U.S. Atomic Energy Commission, will contain author, subject, and report number indexes in each issue. Prompt publication of these issue indexes and of quarterly cumulations of them has been made possible by recent mechanical developments that aid in compiling and publishing indexes. The bibliographic tools, reviews and report services provided by the Atomic Energy Commission demonstrate that the agency recognizes its obligation to make information available in a way in which it can be economically and efficiently used. Some consideration should be given at this Conference to the support and assistance needed in order to make information available.
Mr. Julius Frome concurred with Dr. Libby’s remarks on the tendency to make systems overly complicated. In the U.S. Patent Office persons have worried because a certain mechanized system is capable of handling 12 to 50 interrelated factors, whereas the average person using the system has wanted only three or four. Perhaps scientists using such systems should be trained to make better use of them, or perhaps the systems should not be complicated. It is evident in many instances that the average scientist does not want to use a retrieval system as complicated as might be developed.
Dr. Ralph R.Shaw stated that it is not impossible for abstracting journals to publish indexes promptly. It is only a matter of the amount of effort and money we put into such work. This problem is probably simpler than we
make it. Scientific communication support has not kept pace with the support of research. Perhaps all we need to do to solve our difficulties is to correct this deficiency.
Sir Alfred Egerton began his comments with a few general remarks. Although there were doubts at the time of the Royal Society Conference in 1948, very few persons now regret that the machinery of publication and dissemination of scientific information was inspected at that time. We can perhaps regret that more was not done to follow up that inspection with vigor and insistence on the needed improvements. It has been wise to place first in this Conference the needs of scientists and technologists. Of the total information extant, only part is in the literature. Much of it is stored in the many brains of scientists and technologists. We are as dependent on biological storage as on mechanical or library storage. The mental processes of abstracting and distillation of scientific literature are of great importance to the advancement of science. The greater the mind that carries out those processes, the more chance there is of effective results, though it is true that certain types of minds have a special facility for such work. The needs of scientists working as a team are different from those of the individual scientist. The latter needs special services that provide up-to-date information in his field of interest and needs copies quickly for his own use of papers he wants to study in detail.
Sir Alfred remarked that the study of the use of scientific literature and reference sources in Denmark and Finland by Miss Tornudd contains a great many detailed findings which indicate a need for the following: easily filed abstracts appearing with the original papers in periodicals; formal instruction in library use on the undergraduate level; international abstract services in specialized fields; reduction of the number of different abstract journals; more good critical reviews; and stricter editorial policies to prevent the rehashing of old material. It was suggested that a study of use of the literature in similar institutions with different library services might be made. Miss Tornudd’s study is of interest because it concerns needs in smaller countries where communication by means of the literature is more frequent.
Sir Alfred went on to consider two other papers briefly. The results of a survey of the use of technical literature by industrial technologists, reported by Mr. Scott, indicated that literature is read mainly for general interest and for keeping up to date, rather than for information needed in research. The technologist must be given what he needs rather than what he asks for. Consequently the editor is very important and a key person in the whole information picture. Mr. Herner, in an interview survey of medical scientists concerning use of the literature, found among other things that personal advice from colleagues and face-to-face communication are the most outstanding sources of
information and ideas. All of these papers show the importance of the scientist and technologist as a human being with a brain that can be assisted and fertilized by other brains, as well as by the accumulated stock of knowledge in the literature.
Dr. Shaw expressed the view that most of what we hear at meetings of this sort, his own contributions not excepted, is reminiscent of the lawyer addressing the jury who said, “These are the conclusions on which I base my facts.” The Hogg-Smith paper does not do this, but it does not come to any conclusions. His question then is, “So what?” What good do these studies do if they don’t tell us what to do after we finish them? Most papers on this subject are made “scientific” by pseudo-scientific statistical windowdressing based on inadequate data. Moreover, scientists are frequently asked to give opinions in areas they really have not studied. It is even worse to ask them to grade their opinions. Dr. Shaw suggested the following questions for discussion:
Are there any conclusions to be drawn from these papers that will allow us to take action?
Do we not talk too much about operational efficiency and statistics when we should be considering what can be done in a qualitative way to help geniuses and creative workers?
Are we assuming that because a man is a great scientist he should be an authority on scientific information services?
Perhaps a contribution of this meeting would be to declare a moratorium on publication of inconclusive summaries of incompetent opinions based on inadequate samples.
Dr. I.H.Hogg agreed with everything Dr. Shaw had said. He went on to say that he and Dr. Smith had undertaken their study because it appeared that the scientists in the Industrial Group of the United Kingdom Atomic Energy Authority were not in fact using the services provided for them. The study did serve a purpose in that it persuaded the authorities that the libraries were much too small and that more Russian translations were needed. Conclusions from the study were not available in time to be included in the conference paper. The conclusions and recommendations from the study have since been prepared and they cover 11 pages.
Mr. Saul Herner commented that although there are many things wrong with the studies that have been done in this area, as there are with all studies, they have contributed something. A start must be made somewhere. An imperfect study is a contribution for it gives a point of departure and a basis for comparison. Everyone has said that we don’t know enough from the
studies that have been done. No intelligent critical review of these studies has been made. They show a tremendous amount of consistency in some aspects, such as the temporal span of the use of literature. Mr. Herner suggested that we act on the things we do know and put them to test. We will then find out more about what we do not know.
Dr. Herman H.Fussler commented first on Mr. Fishenden’s report of a study at the Atomic Energy Installation at Harwell. With respect to the use of volunteer subjects, social scientists might express some concern over the possibility of bias. The paper does not indicate the basis on which the validity of the sample was determined. Broadly speaking, the study seems to confirm others that have recognized the importance attached by scientists to current periodicals as a source of current information, and indicates the scattered and semi-fortuitous use of other techniques and services.
Dr. Stephen H.Spurr in his paper presented a personal account of some of the characteristics of research literature in the field of forestry, and recommended that forestry institutions receive a classified subject file of abstracts and citations in card form; also that specialists in the field of forestry participate in the abstracting and classification of papers. Dr. Fussler remarked that such a technique is clearly feasible if the essential conditions to such a bibliographical approach can be met, namely that a central agency compile the abstracts, or take them from the other sources without impairing the other sources economically, and that there be sufficient funds to sustain this rather expensive form of publication. Card techniques seem most likely to work where there are a relatively small literature and a relatively small group of investigators.
Dr. Fussler turned next to Dr. Urquhart’s paper, which presents data relating to the loans of serial publications in 1956 from the Science Museum Library. The data are presented in a terse and abridged form that makes detailed interpretation of some of the findings difficult. The usual concentration of use of scientific serials is strongly confirmed: 80 per cent of the total number of loans is apparently concentrated in less than 10 per cent of the serial titles; of the 9,120 current serial publications received, 4,821 were not borrowed at all; and the most widely held serials are also those most frequently borrowed. Frequency of use, of course, is not the same as a qualitative measure of importance. Dr. Urquhart indicates that in his judgment such data will be useful in shaping; any systematic plan for greater accessibility to scientific serials in the United Kingdom.
The papers in Area 1 had suggested to Dr. Fussler some broad generalizations or observations as follows:
It is evident that the communication of scientific information is a complex and variable matter. It is not evident that our knowledge of these variations is yet as complete as it may need to be.
It is evident that different techniques and channels are used for different needs and that some are more effective than others.
Scientists are likely to learn of major basic developments in their own fields of specialization rather quickly and easily. There is less assurance concerning peripheral information within a specialized field and still less concerning relevant information from other fields.
There is rather substantial evidence that scientists are not in many cases notably systematic in covering the literature, and indeed the existing tools are not exploited by users to anything like their full potential.
Chance associations of ideas and information appear to be important. If this is true, we should be giving thought to information handling devices that would increase the probability of such events.
The studies thus far do not advance an entirely satisfactory explanation of the apparent unsystematic use of available sources and services.
The study of the behavior of scientists in relation to their literature need not be pressed to the last unknown. To do so would require research techniques infinitely more sophisticated than those employed thus far. It should be possible soon, if not now, to begin to outline the specifications for a series of feasible information systems. The specifications should take into consideration three major elements: the habits and needs of scientists as broadly conceived, the economic and material resources that affect the distribution of information, and the characteristics of scientific literature itself.
Mr. R.M.Fishenden said that he believed the results of his survey at Harwell give a fair picture of what is going on and of the failures in the communication system. One general point made by nearly all the papers is that hardly any research worker has access to all the potentially useful information that an information worker may think he ought to have. What we don’t know is whether it matters. It seems that we should direct attention to this question next. What is the penalty, for example, for not having access to the Russian literature or for not bothering to read through all the abstract journals and lists available?
Dr. D.J.Urquhart explained that statistics on the loans of scientific periodicals were compiled for a very practical reason: a library is being built and there
was need to know how many copies of publications to acquire. He would like to receive the results of any other use studies that would help in deciding what publications to acquire for the new National Lending Library of Science and Technology in England.
Dr. Urquhart indicated that this business of measuring what scientists do should not be overdone. The Department of Scientific and Industrial Research in London has made a number of studies and has published some results, but it appears now that the least important results have been published. Whenever a measurement determined policy, it was acted upon and there was no need to publish it. For many purposes crude, simple measurements can be quite useful. We do not necessarily want highly developed experimental techniques. Even crude questionnaires have produced results that have been accepted and acted upon.
The problem of language is a subject Dr. Urquhart thought the conference might seriously consider. It is fortunate that a great deal of the scientific information happens to be in the English language, but this is an accident of history. There is nothing permanent about it. We should seriously consider what language, real or artificial, we can agree to have as a secondary means of communication throughout the world, a language we should all learn.
Dr. Eric de Grolier stated that the results of a questionnaire survey of 200 scientists in France support very strongly the belief that we have a situation of over-communication and that scientists see a need for stricter editorial policies in order to eliminate papers that are not useful. The results also suggest that scientists would favor a drastic reduction in the number of different abstracting services. A clear tendency toward the development of more personal contacts among scientists is discernible. The full report of the study will be published by the French Union of Documentation.
Mr. Herbert Ohlman suggested that since it is difficult for people to remember what they do and use, perhaps we need a system of tallies on the actual pages of materials used, which could be summarized at the time a library makes its inventory. We might even encourage marginal comments in the publications.
Mr. Eugene Garfield expressed the view that too many people are afraid to have opinions. They have to conduct extensive surveys before they are willing to express opinions. If John Shaw Billings had conducted a survey before he went ahead with the Index Catalogue of the Library of the Surgeon General’s Office, we might never have seen it. The phrase “information requirements” is deceiving. It is confused with the information-gathering habits of scientists. The papers in this area do not cover information requirements as such. Mr.
Garfield’s firm recently did a very small inquiry of about five organic chemists and learned that they need and want specific information about the new organic chemicals that are being made.
Dr. Leake commented that scientific productivity is a creative process, which depends upon many factors opposed to stereotyping. Scientists should devise some way of keeping abreast of important developments that occur in fields other than their own and should make some provision for inspiration from either the unusual or the unknown.
Dr. Morse, in bringing the discussion to a close, observed that it is a little surprising that there has not been a review or study of the role of the scientific meeting in the distribution of information. Perhaps such a study should be made. The conclusions in the studies made to date have usually been limited to specific problems. General conclusions of wider interest are rather more difficult and will result from further studies that will be and should be made.
HELEN L.BROWNSON, Rapporteur
PHILIP M.MORSE, Discussion Panel Chairman