Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 163
--> Methods by Which Research Workers Find Information R.M.FISHENDEN ABSTRACT. A survey has been made at the Atomic Energy Research Establishment, Harwell, to discover the methods by which research workers obtain the information they use and read. The object of the survey was to find which methods were most effective in bringing information to their notice, and so to improve the information services in the establishment. The survey was made by two methods: diary cards and personal interview. The results showed that the following were the principal ways by which information was found. The figures represent percentages of all items recorded in the diary survey: regular reading of the current literature, including new reports, 29%; papers found through references in other papers, 9%; personal recommendation, 11%; and scanning lists of titles included in the report lists and information bulletin issued by library, 17%; Nuclear Science Abstracts, 7%; found for readers by the Library, 4%. For the retrieval of old information (22% of all items recorded) there was a marked reliance on personal indexes (4%) and “previous use” (i.e., memory) (10%). All other retrieval methods combined accounted for only 8% of items. There is a strong inference that inadequate attention is paid to systematic searching of the literature and that greater use could be made of library services for such searches. The use of the foreign language literature was small (5%) as was the use of reviews (4%). Comparison with other records, comparison between the diary and interview surveys, and the general consistency of the figures indicates that the results of the diary survey were unexpectedly reliable. An important conclusion is that useful results can be obtained from a much simpler diary card than those used in some previous investigations. The detailed results, relating as they do to a particular set of circumstances, are of limited general interest, but they give valuable and much needed guidance on the ways in which the AERE information services should be developed. R.M.FISHENDEN Atomic Energy Research Establishment, Harwell, England.
OCR for page 164
--> 1. Summary of services given by A.E.R.E. Library The Atomic Energy Research Establishment at Harwell is the Headquarters of the Research Group of the United Kingdom Atomic Energy Authority. It has a scientific and technical staff of nearly 2000, with several thousand ancillary staff. The Library serves the whole Establishment but for the purposes of this paper only the scientific and technical staff will be considered. The Establishment is divided into fourteen subject Divisions under Division Heads, e.g., Chemical Engineering, Chemistry, Engineering, Isotope, and Reactor, and each Division is broken down into Groups under Group leaders. The library contains some 15,000 books, 12,000 volumes of periodicals, 22,000 pamphlets and 145,000 reports, and receives about 750 current periodicals. It is divided into three sections: (a) the Reading Room (published literature), (b) the classified and unclassified Report Rooms, and (c) the Information Office. The Reading and Report rooms are under qualified librarians, the Information Office is staffed by scientists and the whole is under a librarian with scientific qualifications. 1.1. READING ROOM AND REPORT ROOMS These rooms provide reference and lending services and a photocopying service. They maintain the following catalogues: (a) Author catalogues of books, pamphlets, periodicals, U.K.A.E.A. reports and other reports which do not appear in Nuclear Science Abstracts. (N.S.A. is an abstract journal issued by the U.S. Atomic Energy Commission and is used at Harwell as a catalogue and accessions list of U.S.A.E.C. unclassified reports). (b) Subject catalogue of books, pamphlets, and periodicals. It is classified by the U.D.C. (c) Alphabetical subject catalogue of U.S.A.E.C. unclassified reports. This is provided by the U.S.A.E.C. The subject catalogue of all other reports is held by the Information Office. (d) Serial catalogue (i.e., catalogues under series numbers) of classified and unclassified reports. Accessions lists are issued: (1) recent additions (books, pamphlets, periodicals), monthly; (2) “Unclassified” and “Official Use Only” reports, weekly; (3) “Confidential” reports, fortnightly; (4) “Secret” reports, monthly. U.S.A.E.C. unclassified reports appearing in Nuclear Science Abstracts (N.S.A.) are not listed. N.S.A. is widely distributed about the Establishment. 1.2. THE INFORMATION OFFICE This office provides scientific and technical information and prepares bibliographies in answer to requests. It abstracts and classifies by U.D.C. the reports
OCR for page 165
--> received in the library, other than those in N.S.A.; the abstract cards forming the subject catalogue to the report literature. Other services provided are: A weekly “Information Bulletin” of articles of interest in the current periodicals. Abstracts of the articles are not included. The list is intended as a quick notification of what is available. Notifications to individuals of articles in the current literature of interest to them (notification slips). Despatch of new reports to persons most interested in them. Translation from all languages except French. It is expected that members of the Establishment can read French. 1.3. DIVISIONAL LIBRARIES These supplement the Main Library. Each Division has at least one. They are not subject to the control of the Librarian, but there is close liaison. Most of the material held by Divisional Libraries is also in the Main Library. In addition to the above collections, authors of reports send copies to individuals on the Site by putting them on the initial distribution of the reports concerned. 2. Choice of survey method Two methods were considered for this survey: first, interviews of research staff by a member of the Information Office and second, the diary card method. The expected advantages and disadvantages of the two methods are summarised below. 2.1. DIARY METHOD The advantage of the diary is that it offers, at least in theory, a method of securing an accurate record of the information used by the diary keeper and how the information was obtained. On the other hand, the method is more dependent than is the interview method on cooperation from diary keepers. Furthermore, the record is unlikely to be 100% complete, and omissions may not be random. 2.2. INTERVIEW METHOD This method is much the more flexible, and allows discussion of interesting points raised by those interviewed. The main objections are that it is subjective, and that, since it relies on memory, recent events are likely to receive too much prominence. In discussion prior to the survey, opposing views were strongly held. Critics of the interview method held that so much subjective information was already
OCR for page 166
--> available on information methods that a more exact approach was required, while critics of the diary method held that the filling in of cards would be so erratic that any appearance of numerical accuracy would be illusory. So firmly were these views held that compromise proved impossible, and the decision was taken to conduct independent surveys by the two methods. This had the incidental advantage that the surveys could be planned so as to make some of the results comparable. In the event the results of the two surveys were in fair agreement (see section 6), and the weaknesses of neither method seem to have been as serious as expected by its critics. 3. The Diary Card Investigation The diary card investigation of the methods used by scientists in finding information follows roughly on the lines of similar investigations by Bernal (1) and R.R.Shaw (2). However, a fundamental decision made at the start was that reliable results could be obtained only if record keeping was made very simple: the aim was a single mark on the card for each item recorded. The investigation also differed from those undertaken previously, in that it was carried out in a research establishment with a relatively large information office in the Library. The diary card used, together with the supplementary notes given to diary keepers, are reproduced in Appendices 1 and 2. It will be seen that the card enables the diary keeper to show by a single dot how each useful “unit of information” was found, and whether it was in a report, review, book, or elsewhere. A second dot is required only for foreign language literature. Provision is also made to distinguish between background reading and information used directly on a job. It was expected that this information would be a small proportion of the whole, but that in getting it there would be a comparatively frequent use of some method of retrieval. If the figures were diluted with a larger number of background reading items (not involving retrieval problems), it was thought that it would be more difficult to identify the most effective retrieval methods. In the event, the distinction did not prove useful. [See 4.2 (b).] Of the 17 horizontal lines on the card 1–10 cover information found by staff without library aids, and 11–17 information found with library aids. The final layout of the card was decided after a one month pilot run in which members of the Chemistry Division at A.E.R.E. participated. This pilot run showed up a number of important omissions in the trial card, and also led to the attempt to separate “background” from immediately useful information.
OCR for page 167
--> The design of the diary card, and problems met in using it, are discussed further in Appendix 3. The information on the completed cards was transferred to Hollerith punched cards, together with code numbers to provide the following information: (a) identity of diary keeper, (b) Division, (c) “pure” or “applied” research worker or engineer, (d) Group Leader or junior. The survey was confined to staff in those grades containing only honours graduates. A sample of 80 staff was aimed at, and each Division was asked to provide volunteers pro rata with its size and to choose the volunteers so as to give a representative cross section of graduate staff. The cooperation from Divisions was sufficient to give a reasonably representative sample of the establishment, and 63 completed cards were returned. The diaries were kept for approximately 2 months. 4. Results of diary card survey The results of the diary card survey are presented in tables giving the following information: Table 1, overall results, i.e., totals for each box for all 63 participants; Table 2, number of participants making any use of particular methods; Tables 5 and 6, as for 1, but broken down into separate classes. In addition, the record was discussed with 11 diary keepers. This is referred to subsequently as the “oral sample.” There was surprising unanimity that for written information (columns A-H) the record was 80–90% complete and that in this field it presented a valid picture of the information used. All diary keepers found, however, that it was impracticable to find a logical basis for recording information received orally, or by written private communications. 4.1. OVERALL RESULTS The overall results of the diary card survey are given in Table 1. Because the survey failed for oral information and written private communications, the results for columns I-P of the card have been excluded from the table (but see 4.2 (k) for figures on the use of Conference proceedings). The explanatory notes reproduced as Appendix 2 will be found helpful in interpreting the results. The results in Table 1 are supplemented by Table 2, in which is shown the number of diary keepers making any positive use during the survey period of a particular service or information tool. 4.2. DISCUSSION OF OVERALL RESULTS (a) Reading of the current literature is as expected the most important single method by which information is acquired. However, the results emphasize
OCR for page 168
--> TABLE 1 Overall results (63 diarists)
OCR for page 169
--> TABLE 2 Proportion of diary keepers making any positive use of a particular source or kind of information Nuclear Science Abstracts 32% Chemical Abstracts 17% Physics Abstracts 17% Other abstract journals 19% Private index 45% Information bulletin 37% Report lists 50% Library notification slips 20% Reviews 41% Foreign literature 30% that the current report literature is not read in the same way as current periodicals. Unless a reader is sent a personal copy of a report (row 3 of the diary card) he has to make his own selection of titles, and he does this principally from Nuclear Science Abstracts (row 5) and the report lists issued by the library (row 14). (b) Although the number of items found from references in another paper is substantial, it was expected that this would be a relatively more important source. An independent survey which showed very limited cross-referencing between papers in a particular subject field, may provide the explanation. (It is hoped to report the results of this survey to the Conference.) (c) Comparatively little use is made of abstract journals. Only one-third of the diary keepers made any positive (i.e., successful) use of Nuclear Science Abstracts, in spite of the fact that N.S.A. is widely distributed and is the only method of notifying staff of U.S. unclassified reports received (apart from the limited library service covered in rows 16 and 17). The sample oral enquiry confirms this conclusion. (d) Twenty-eight (45%) of the diary keepers show some use of a private index, and it is doubtful whether the row 9 total reflects adequately their importance. Of the oral sample 8 had shown no use of a private index, but of these 5 maintained one and 2 said they had forgotten to enter its use in the diary. There is no doubt that these indexes represent a valuable source of information of which more use might be made if the owners were willing to receive other people’s enquiries. (e) Of the services offered by the Library, much the most widely used are the information bulletin and the report lists. 40% and 50% respectively of the diary keepers showed some positive use of these lists and the oral sample indicated that about half the remainder were looking at these lists, but had found nothing of interest during the period. Altogether an estimated 70% of diary keepers were using each list. (f) The use made of other library services (rows 11, 12, 16, and 17) was
OCR for page 170
--> numerically comparatively small. This might be expected, as library indexes would only be used (or a request made to library staff) when other means had failed. The notification slips are intended to cover articles in journals which a worker would not be likely to see himself: a high proportion of the slips sent out were used (see Section 6). However there is a strong indication that inadequate use is made of library resources for literature searches (see h). (g) Row 15 covers only the use of bibliographies already in existence. No bibliography was prepared for a diary keeper during the survey period. (h) Most effective retrieval methods. Of the methods used for finding information, it is possible to make a rough division into those applicable to the current literature, and those applicable to retrieval of older material, as follows: (i) For current literature Rows 2, 3, 5, 13, 14, 16, 17 (ii) For retrieval Rows 6, 7, 8, 9, 10, 11, 12, 15 (iii) Not allocated Rows 1, 4 Nuclear Science Abstracts is included in (i) since it is used principally as a list of the current U.S. report literature. The numbers of items obtained by the different retrieval methods are shown separately in Table 3. It will be seen that two retrieval methods, private index TABLE 3 Most effective retrieval methods Retrieval method Items % Abstract journals 98 23 Private index 83 19 Previous use 186 43 I found it in library indexes 26 6 Library found it for me 25 6 Bibliography 11 3 Total 429 100 and previous use, account for nearly two-thirds of the total. It seems a reasonable inference that the extent to which careful literature searches are made is small, and that greater use of library services for this purpose could profitably be made. Incidentally, the separation of the record into “background reading” and “use” items did not assist in identifying the most effective retrieval methods. However, as might be expected, three-quarters of the “retrieved” items were shown in the “use” columns and only one-fourth as background reading. (i) The use made of the foreign language literature was slight, roughly 5% of items recorded being from this source. [This is the same result as obtained by Shaw (2)]. Less than half the diary keepers used the foreign language
OCR for page 171
--> literature at all. This is a not unexpected result, but it demonstrates once more how little foreign language information is used. It would be interesting to try and assess the cost of overlooking the remainder. (j) Although the distinction between “background reading” and information required for use on the job is not precise, the results indicate that reading is spread roughly equally between the two. Exceptions are that of items read in current periodicals, about two-thirds are background, whereas books are read only one-third for background. (k) The use made of reviews as a source of information was unexpectedly small, about 4% of items coming from this source. Less than half the diary keepers used reviews at all. An explanation may be that in a relatively new field only a small proportion of the literature is included in reviews, but the result does not seem to bear out the views expressed at the 1948 Royal Society Conference. The use of Conference Proceedings slightly exceeded the use of reviews. (These are the columns I/P results, not included in the tables.) (l) 1896 items of information were recorded by the 63 diary keepers, an average of almost exactly 30 each. (This excludes 262 items in columns I-P of the card, which were excluded from the results for the reasons given in Section 4.1.) There was a considerable spread between a minimum of 3 and maximum of 103. The distribution of items recorded per diary keeper is shown in Table 4. TABLE 4 Spread of number of items recorded per diary keeper Items recorded Diarists in range 0–10 13 11–20 16 21–30 10 31–40 10 41–50 3 51–60 3 61–70 3 71–80 1 81–90 2 91–100 1 101–110 1 Total 63 It will be seen from the table that 5 diarists recorded over 70 items each (in the range 77–103). All the rest recorded less than 65. If the 5 high scorers are omitted, the average for the rest drops from 30 to 25. If allowance is made for Christmas holidays, the average number of recorded reading acts per week is about 4.3, somewhat lower than Shaw’s figure of 5.8 (2).
OCR for page 172
--> 4.3. BREAKDOWN OF RESULTS FOR DIFFERENT TYPES OF DIARIST As mentioned earlier, the Hollerith card records were coded so that results could be totalled separately for senior and junior staff, and for workers in pure and applied research. The results are shown for columns A-H only as there was no significant difference in the use of the foreign language literature. In addition, the figures for “background reading” and “use” are combined as there is no significant difference in the proportion of each between the broken-down results and the overall results. Comparison of senior and junior staff. The results for senior and junior staff are compared in Table 5. The similarity between the results for senior and junior staff are remarkable. In the four columns A/B—G/H the greatest difference is 2%, (and the similarity is maintained when the 8 columns A-H are added separately). There are slightly greater differences in the methods used to find information. Senior staff receive more reports on initial distribution (as would be expected) and rely more on references in other papers. Junior staff rely more on previous use and abstract journals, and are slightly more willing to use library indexes (catalogues). The sample contained 23 senior staff of roughly group leader status and 40 junior staff, the average number of items recorded being 27 for senior staff and 32 for junior. The consistency of these figures encourages the belief that some reliance can be placed on the results of the survey. Comparison of results for pure and applied research workers. The results for pure and applied research workers are shown in Table 6. As might be expected, there are considerable differences between them. In particular the former make practically no use of the report literature, over 70% of their reading being in journals, whereas the latter make roughly equal use of the report literature and journals. This difference is reflected in relative use made of different sources of information. The “pure” group make noticeably more use of references in other papers, but surprisingly little use of abstract journals. The “applied” group make greater use of library services. The sample contained 19 pure research workers and 34 applied, the balance being engineers or not easily allocated to any group. The proportion of pure research workers exceeds the establishment average (about 20%) because of the inclusion of 6 staff of the Medical Research Council. The average number of items recorded was 25 for pure and 37 for applied research workers. It should be noted that the Medical Research Council staff have an independent Library of their own, and make very little use of the services covered by rows 11–17 of the diary card.
OCR for page 173
--> TABLE 5 Comparison of senior and junior staffs
OCR for page 174
--> TABLE 6 Comparison of pure and applied research workers
OCR for page 175
--> 5. The interview survey A representative sample of 50 graduate members of staff were selected at random for the interview survey. Each of the selected subjects was interviewed by Mr. B.Wilson, a member of the Information Office staff. The list of topics discussed at each interview is reproduced as Appendix 4. The interview survey was designed primarily to obtain information of direct value to the Information Office in planning their services. Only those results bearing on the main theme of this paper are presented here, in the form of a comparison with corresponding results from the diary survey. The results of Mr. Wilson’s survey are being written up separately as an A.E.R.E. Library report and I have to thank him for permission to quote these extracts from them. 6. Comparison of diary and interview surveys Both surveys give a figure for the proportion making any use of a particular reference source, and this has been chosen as a basis for comparison. The best comparison is possible in the case of the Information Bulletin, the Reports lists, and Nuclear Science Abstracts. For these the interviews gave figures for regular and occasional use. The diary gave figures for any positive use in the survey period and the oral sample indicated whether the remainder did not use the reference source, or had found nothing in it during the survey period. The figures, together with rather less complete figures for chemical abstracts and private indexes are included in Table 7. TABLE 7 Proportion of sample making use of particular reference source Interview Diary Source Regular use Occasional use Total Positive use in survey period Estimate of others who use Approx. Total Information Bulletin 60% 6% 66% 37% 31% 70% Report list 52% 6% 58% 50% 16% 65% Nuclear Science Abstracts 40% 16% 56% 32% 17% 50% Chemical Abstracts 22% 2% 24% 17% — — Private Index — — >60% 45% 30% 75% The agreement between these figures is within the probable errors. A cruder comparison is possible between the results from the two surveys for the use made of bibliographies, of notification slips, and of reports issued on library initiative (rows 15, 16, and 17 of the diary card). Here, all that is available is a
OCR for page 176
--> “% using” figure from the interviews and a “positive use” figure from the diary cards. Interview “Any use” % Diary positive use Notification slips 54% 20% Reports issued on library initiative 48% 24% Library bibliographies 18% 12% It was possible in the case of the notification slips to check the diary card results with library records. These showed that slips were sent to 13 of the diary keepers, and 13 showed a positive use of the service—a remarkable correlation. A total of 49 slips were sent out, and the diary records showed 39 items found from slips. This confirms that the records are reasonably complete, as recipients would not follow up all slips received. The reason that the diary card figures are in each case lower than the interview figures is that they cover a shorter period. It is known that at least 44% of those interviewed were sent slips in the last few months, which is in reasonable agreement with the “any use” figure of 54%. Although the interviews gave no figures comparable with the diary survey for the use of foreign literature, they confirmed that use was slight, and indicated in addition that important Russian and German material is not being used because staff do not request translations. REFERENCES 1. BERNAL, J.D. The Royal Society Scientific Information Conference Report, Paper 46, 1948. 2. SHAW, R.R. Pilot study on the use of scientific literature by scientists. 1956.
OCR for page 177
--> APPENDIX 1
OCR for page 178
--> APPENDIX 2 Notes on use of diary cards Make a single mark (preferably a dot) on the card for each useful “unit of information” obtained. The card is laid out so that you can indicate by this dot how you found the information (e.g. from an abstract journal, by receiving it on initial distribution, or by personal recommendation), what the information was in (e.g., a report, a journal, a review) and whether the information was for background reading (B/R) or direct application to a job (use). E.g., for an article found in the information bulletin of direct application to your job, put a dot in box D13. “Unit of information” should normally be interpreted as a report, article, review, lecture, but make a record even if only part of the information in a report etc. is used. For foreign publications make an additional dot in column Q or R as appropriate. It is the intention that important information obtained in discussions (i.e., oral private communications) should be recorded. Boxes K4 and L4 will normally be appropriate. If information reaches you several ways enter the first only. If you come across information of a kind or by a method not provided for on the form, e.g., newspaper article, or pamphlet, don’t record it. A pilot run with spare boxes for such items showed they were insignificant in number. For borderline items between “background” and “direct application” items, make a dot on the borderline between the two relevant boxes. There may sometimes be two stages in finding information, e.g. personal recommendation (have you seen that Brookhaven report by Smith and Brown) followed by identification in an index. If so, put a cross in both boxes. If during the period you are keeping this diary you undertake a special literature survey e.g., because you are starting a new project or writing a report, please use a separate diary card for this. Disregard the numbers written sideways. These are for use when results are transferred to Hollerith cards. The validity of your record will not be affected by an occasional omission (provided these omissions are random!), so please don’t give up because you forget to use the card for a few days. If one of the boxes on your card gets full, start another card. The Divisional representative who issued your card has a stock of spares. R.M.FISHENDEN 12th November, 1957.
OCR for page 179
--> APPENDIX 3 Discussion of diary card layout In general the card layout appears to have been quite satisfactory, but the following points may be of interest to anyone attempting a similar survey: “Unit of information” is rather a vague term on which to base an investigation. Nevertheless, it is difficult to improve on. It appears to have been a satisfactory definition for written information. Only 2 people said the definition was so vague they could not keep a satisfactory diary, although a few others may have failed to keep a record for the same reason. The diary card failed to offer a practicable method of recording information received by the spoken word. No one could find a logical basis on which to decide what to record. The card makes no provision for recording useless information received. One or two people mentioned this, but it is doubtful whether the extra complication would be worth while. The distinction between “background reading” and “use” did not assist very much in identifying the most important retrieval tools. A better alternative might be to distinguish between current literature and “old” information. The card did not distinguish between information obtained (i) by attending lectures or conferences and (ii) by reading a written record of the proceedings. APPENDIX 4 Topics discussed at interviews Rank, division, position, length of service at Harwell. Extent of use of (a) Library services and (b) Information Office services. Proportion of use devoted to background reading and to specific searching. Frequency of use of the following, and comments on their usefulness: Information Bulletin; accessions Lists (including reports and books); bibliographies; notification slips from daily scrutiny of journals; initial circulation of reports by the Information Office. Frequency of consultation of the following: journals; books; reports; Nuclear Science Abstracts; other abstract journals; U.S.A.E.C.’s subject catalogue of unclassified reports; library author catalogue; Information Office (security classified); subject catalogue of reports; library subject catalogue; press cuttings; lantern slide catalogues; Microcards. Extent of use of (a) Library staff, and (b) Information Office staff. Personal indexes. Use of foreign language material. Journals on subscription and journals read or scanned regularly. Questions were also asked on a number of topics of domestic interest to the Library, e.g., on the efficiency of certain services.
OCR for page 180
--> This page intentionally left blank.
Representative terms from entire chapter: