National Academies Press: OpenBook

Proceedings of the International Conference on Scientific Information: Two Volumes (1959)

Chapter: An Operations Research Study of the Dissemination of Scientific Information

« Previous: The Transmission of Scientific Information
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

An Operations Research Study of the Dissemination of Scientific Information

MICHAEL H.HALBERT and RUSSELL L.ACKOFF

The research reported here is still in progress at Case Institute of Technology under the sponsorship of the Office of Scientific Information of the National Science Foundation. Between this writing and the International Conference on Scientific Information, work on this project will continue. Only preliminary results are available at this time. The additional results obtained in the next six months will be reported at the time of the International Conference.

General design of the research

The question originally asked of the Operations Research Group at Case by the Office of Scientific Information was: What is the possibility of applying Operations Research to problems in the dissemination of recorded information? The research reported here is a partial answer to this question. To understand its development it is helpful to be aware of two essential characteristics of Operations Research. The first is that Operations Research is concerned with the application of scientific method to the study of systems of organized activity rather than to the components of such activity. Its orientation is “whole-istic.” Secondly, Operations Research is operationally oriented. This means that its primary concern is with affecting the way systems operate and not merely in providing “interesting information.” In brief, it seeks to provide a basis for effective action.

With this much in mind we can reconstruct the logical development of this project.

The system of scientific communication can be thought of as having three phases: (1) production, the formulation of a message which may take the form of an article, book, memorandum, speech, conversation, etc.; (2) distribution,

MICHAEL H.HALBERT and RUSSELL L.ACKOFF Operations Research Group, Case Institute of Technology, Cleveland, Ohio.

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

the dissemination of the message as by publication, presentation at a meeting, etc.; (3) consumption, listening to or reading the message.

Of these three types of activity the most organized is the phase of distribution since this involves the activities of scientific societies, research institutes, publishers, libraries, and related organizations. Production and consumption of information is primarily individualistic and independent activity which, because of its lack of organization, is the most difficult to manipulate or control in any way. Consequently, the attention of this research was concentrated on the distribution phase of the communication process, but this concentration involved a concern with the interaction of the other phases with distribution. That is, it was decided that improvements in scientific communication could best be obtained by manipulation of distribution, but this did not mean that production and consumption could be ignored. To the contrary, changes in distribution could only be evaluated in terms of their effect on production and consumption.

What aspects of distribution should be studied? Consideration of the activities of organizations engaged in the dissemination of scientific information led us to conclude that their primary problem involved the question of how their resources should be used. Their resources include not only money, but also men, material, and machines. That is, how should the resources of these organizations be allocated to the various ways of disseminating scientific information so as to maximize the effectiveness of the communication system?

To answer such a question a measure of effectiveness of the system is required. It seemed clear that the system is concerned with increasing scientific productivity. It seemed equally clear that an acceptable measure of scientific productivity was not likely to be obtained within the time available for this project. Nevertheless, it is of value from the viewpoint of research design to consider how an answer to the problem of resource allocation would be obtained if such a measure were available.

Ideally, we would first like to identify all the alternative ways in which the organizations considered could invest their resources. For example, they might increase their abstracting services; they might initiate digesting or reviewing services, or they might develop new cataloguing systems. These are but a few of the many alternative types of activities in which these organizations could engage. Ideally we would like to identify all of them.

Secondly, for each possible alternative we would like to determine how any specified investment could best be used. For example, if, say, $100,000 were to be made available for improving abstracts, how could this sum be used so as to maximize the resultant increase in scientific productivity. Put more generally, we would like to generate a “pay-off” function for investments of various

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

amounts in each alternative activity, assuming these investments were used in the most effective way.

If the alternatives could be completely identified, and if the best use of an investment in each could be specified, and if a pay-off function could be derived for each alternative, then we could construct what is called an “allocation model” of the system. Methods are available, once such a model is constructed, for determining for any amount to be invested, how best to invest it; and, more generally, how much ought to be invested and how.

Such thinking, as we have indicated, is highly idealized, but it provides a standard against which a practical research design can be developed. We would like to come as close to the ideal as possible. The practical adjustments that are required arise primarily out of two difficulties. The first we have already mentioned: the difficulty of developing an acceptable measure of scientific productivity. The second difficulty arises out of the very large number of alternative courses of action which are available to disseminating organizations.

Since we could not expect to measure scientific productivity directly we sought an aspect of scientific activity which (1) could be measured objectively and (2) if increased, would also increase scientific productivity. The time available for scientific research is such an aspect of scientific activity.

An effort which attempts to maximize the amount of time available for scientific research has several advantages. First, it is widely recognized that we are suffering from a considerable shortage of scientific manpower. Consequently, any activity which promises to make more research time available with the same number of men may help to solve this shortage problem. Secondly, the amount of time available for scientific research can be measured objectively.

It should be noticed that, although there is an assumption implicit in the use of “time available for scientific research” as a measure of effectiveness to the effect that if this time is increased productivity will also increase, there is no assumption made about the nature of this relationship, that is, its mathematical structure. It does not follow from this assumption, for example, that doubling the time available for scientific research will double scientific productivity. All that is assumed is that there will be some increase in productivity.

Because of the large number of alternative courses of action that could be considered, it is necessary to have a preliminary screening which will reduce this number to a manageable magnitude. If we knew how scientists actually spend their time now, then we could concentrate on alternatives which are most likely to affect significant portions of that time. Consequently the practical research design was conceived of as addressing itself to three questions in sequence: (1) How do scientists actually spend their time? (2) In what types

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

of scientific activity are there the greatest potentialities for reducing time expended without reducing scientific output? (3) How can these potential reductions be realized in the most effective way?

At the time of this writing (1) has been answered for a specific science and (2) is answered in part. The third question has not yet been considered. By the time of the conference it is hoped that (2) will have been completed and that significant progress in (3) will have been made.

How scientists spend their time

In order to study the way scientists spend their time it is necessary to have a classification of such expenditures. The classification is critical. For example, one might observe what portion of their time scientists spend wearing jackets, but this is not likely to indicate fruitful possibilities for saving time. We need a way of looking at time expenditures which suggest things we can do to affect these expenditures.

First, it was clear that we wanted to distinguish between time spent on scientific activity and time spent on other types of activity. It was also clear that we wanted to distinguish between time spent in communication and time spent in other activities. This yielded a basic fourfold classification. It is also necessary to know how time spent on non-communicative scientific activity is allocated. This type of activity seemed to be conveniently classifiable into: thinking or planning alone, setting up or maintaining equipment, using equipment to generate data, and treating data.

Nonscientific activity can be divided into business activity and that which is personal and social.

These considerations led to the following basic classification of activities of scientists:

Scientific communication

Non-scientific business communication

Thinking or planning alone

Equipment set-up and maintenance

Equipment use

Data treatment

Personal and social

None of these

Out of area

“None of these” is included to cover any miscellaneous activities that cannot otherwise be classified. Discussion of the observational plan in the sequel will show why “Out of area” was required. Detailed definitions of these categories will be found in Appendix 2.

Since our primary concern is with scientific communication, this category had to be further broken down. This breakdown is essentially concerned with (a) the phase of communication: sending, retransmitting, and receiving, (b)

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

who is involved in the communication, and (c) the channel or medium of communication employed.

Space does not permit detailed justification for the classification of these three phases of communication, but, fortunately, the classification, we think, tends to justify itself. The phases of scientific communication were broken down as follows:

  1. Hearing question

  2. Reading question

  3. Reading for use

  4. Reading for general information

  5. Hearing information

  6. Working out material

  7. Editing material received

  8. Writing information

  9. Telling information

  10. Writing question

  11. Asking question

  12. General discussion

  13. Discussion about a received communication

  14. Reading for retransmittal

  15. None of these

The number of the following types of persons involved in each communication was sought:

  1. Mathematicians, statisticians

  2. Physical scientists (and engineers other than chemical)

  3. Chemists (and chemical engineers)

  4. Biologists and medical men

  5. Behavioral scientists

  6. Secretaries, technicians, etc.

  7. Other personnel in the same organization

  8. None of these

The channels involved were classified as follows:

  1. Oral

  2. Unpublished written

  3. Book

  4. Article

  5. Abstract or review

  6. None of these

Explanations of each of these categories can also be found in Appendix 2. The observational forms themselves will be found in Appendix 1. But these will be better understood after a discussion of the remainder of the research design.

Several different methods of collecting information on how scientists spend their time were considered and tested in the field. An effort was made to develop a procedure which would yield reliable, accurate and objective data. These requirements dictated that the report made on the scientist should be prepared by someone other than the scientist himself. The classification of activities lends itself to use by an observer. Only a few categories required any consultation with the scientists themselves, and these involved only questions of fact, not a solicitation of opinions or attitudes. The pretests supported these contentions. In order to get a representative sample of time the “ratio delay”

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

method was adapted to our purposes. This consists of observing the scientist at randomly selected moments of time. From observations so made estimates can be made of the over-all allocation of time and the accuracy of these estimates can also be determined.

After extensive pretesting and collection of information as to the time required in making observations and in setting up to make observations, and after estimating the number of observations required to obtain results of acceptable reliability, it became clear that the resources available for this study would permit analysis of only one scientific discipline. The field of chemistry was chosen primarily for two reasons:

  1. Chemists constitute the largest scientific group, numbering approximately 80,000.

  2. Among scientific groups the chemists have probably been among the most concerned with problems involving the dissemination of recorded scientific information.

It was considered essential to select the chemists to be observed in such a way that inferences could be drawn from the resultant data to as large a portion of the total population of chemists as possible. It was our feeling that in too many studies of scientific communication groups are selected for their convenience and consequently provide no scientific basis for generalization of results obtained. Consequently, with the cooperation of the American Chemical Society and the National Science Foundation a proportionate stratified systematic random sample was drawn from the population of chemists in the 150 metropolitan areas of the United States. Details of the sampling procedure are given in Appendix 3.

The sample yielded 50 observed groups of chemists, employed by 45 different industrial organizations and 5 groups of chemists in universities. This led to approximately 25,000 observations of about 1500 chemists. This preliminary report deals with data on the industrial groups only (approximately 18,000 observations).

Each chemist was observed at two random moments of time each day for 9 consecutive days. Observers were company or university personnel trained and carefully supervised by members of Case’s research team. Using “local” personnel for observers considerably reduced the amount of questioning required of the subjects in the observational process. Observations were recorded on the two forms shown in Appendix 1.

Since some refusals by organizations to cooperate was anticipated, a larger sample than was required was selected. Four companies refused to cooperate. This could, of course, introduce some bias in the results if these companies have

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

different characteristics from that of the companies which cooperated. The closest examination we could make for such differences failed to reveal any such characteristics. The reasons for refusal were quite reasonable in most cases and failed to be related to the characteristics of the company as far as we could determine.

So much for the discussion of the plan of the research. The remainder of this paper will deal with the results obtained.

Analysis

The question, “How do chemists actually spend their time?” really has three aspects of equal importance. The first can be stated as: What is the average amount of time spent by chemists in each of the activities classified? This can also be stated in terms of the probability that a randomly observed “chemist-moment” will be classified as a particular activity. This probabilistic form of the question is basic to the mathematical analysis to be described later.

The second aspect of the question as to how chemists spend their time is the consideration of the variability among groups of chemists. This aspect also includes the dynamics of time allocation. Here we are interested in any relationships between or among the various classes of activities. This may be stated as: What are the distributions around the averages of time spent, and what functional relationships exist among the averages? One very obvious relationship is that the percentage allocations must add to 100. However, there are other less obvious relationships. This aspect of the time allocation is naturally the most significant in terms of the basic purposes of the study.

These interrelations show where the potential time savings can be obtained. It is not enough to see where the most time is being spent. We must see how the time division among some activities responds to changes in amount spent in other activities. The third aspect of the question concerning the disposition of time by chemists is concerned with the influence of the various environments within which chemists work. Such factors as the type of research in which the work unit is engaged, the availability of scientific literature, salaries, and the professional composition of the work unit may influence time allocations of chemists. More will be said about this point following the discussion of the first two aspects mentioned above.

Environmental data are in the process of being gathered by the questionnaire which appears as Appendix 5 of this report. To date, half of the 50 units included in the study have replied and analysis has just begun. Consequently, there are no preliminary results to report at this time.

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

Time allocation

The results of the 18,000 in area observations show clearly that the conventional picture of the chemist as a man in a white lab smock pouring chemicals into test tubes is a distorted picture. If we lump all experimental, equipment, and data time together we can account for only 36 percent of the chemist’s time, while necessary communication (scientific and business) takes up 44 percent of the chemist’s work day. Figure 1 shows the division of time among communica

FIGURE 1. Chemists’ time allocation summary.

tion, experimentation, and facilitation. The figure also shows the maximum and minimum values for the 100 groups of chemists studied. Most of the following analyses are based on these 100 units which were derived as follows. Each of the 50 groups of chemists was observed 18 times during 9 consecutive days. These 18 observations were divided into “early” (the first 9 observations) and “late” (the last 9) for each group, yielding the 100 units for further analysis.

Figure 2 shows the time division for the eight categories used on form 957–1: Scientific communication; Business communication; Thinking or planning

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

alone; Equipment setup or calibration; Equipment use; Data treatment; Personal and social; and Miscellaneous. The maxima and minima are also shown. It is obvious that Scientific communication is the largest single category. Since our observers were, in most cases, professional chemists and working members

FIGURE 2. Chemists’ time allocation, detail.

of the group being observed, there is little likelihood they would mistake casual conversation (which might contain technical jargon) for Scientific communication. Also, since each observation that was classified as Scientific communication required filling in an additional form (957–2), there was no reason for the observers to err in the direction of overreporting this category. We feel, therefore, that chemists spend at least the reported 33 percent of their time in Scientific communication. Also it is the only category that has a minimum other than zero. The data in Figure 2 are presented in Table 1. Except for the two largest categories, Scientific and Business communication, the maxima are approximately four times the average. With the larger categories this drops to 2 to 3 times the average. This, of course, is due to the limiting effect of an upper bound—100%. Removing the effect of this constraint was one difficulty in the development of the detailed analysis.

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

TABLE 1 Percent time allocation

Activity

Minimum

Average

Maximum

Scientific communication

15.7

33.4

61.4

Business communication

0.8

10.4

40.0

Thinking or planning alone

0.0

6.0

25.6

Equipment setup

0.0

6.2

25.9

Equipment use

0.0

23.4

70.1

Data treatment

0.0

6.4

31.6

Personal and social

0.0

9.8

33.6

Miscellaneous

0.0

4.4

13.6

A further set of time averages was obtained for various aspects of scientific communication. These are shown on Figs. 3, 4, and 5, and in Table 2. The exact definitions of these categories, and all the others used in this analyses are

TABLE 2 Percent scientific communication time allocation

Activity

Minimum

Average

Maximum

Total scientific communication

15.7

33.4

61.4

General discussion

0.0

10.3

35.3

Oral, non-discussion

0.0

9.2

28.0

Total

Written

3.9

14.3

45.0

Unpublished written

0.0

9.5

40.0

Published written

0.0

4.9

18.4

Sending, oral

0.0

4.5

17.7

Receiving, oral

0.0

3.8

19.4

Sending, written

0.0

5.0

15.0

Receiving, written

0.0

7.2

18.4

Retransmittal

0.0

2.7

20.6

Reading articles

0.0

2.6

13.7

Reading for use

0.0

3.9

14.3

Reading for general information

0.0

3.2

18.4

Communication with other scientist, non-chemist

0.0

2.7

16.3

Communication with other company personnel, secretaries, technicians

0.0

7.1

25.7

Communication with chemists

5.3

21.4

54.5

contained in Appendix 2. Some of these relationships are suggestive of experiments, and some lend themselves directly to interpretation. For example, if we construct a fourfold table from Sending oral, Receiving oral, Sending written; Receiving written (Table 3), three things become obvious.

TABLE 3 Sending-receiving vs. written-oral, percent

 

Sending

Receiving

Total

Written

5.0

7.2

12.2

Oral

4.5

3.8

8.3

Total

9.5

11.0

20.5

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

FIGURE 3. Chemists’ communication time, allocation to channel.

First, it is apparent that of the four categories here discussed, more time is devoted to receiving recorded communications than to any of the other three categories. This, of course, is to be expected. Not as expected is the fact that while more time is spent in receiving than sending written material less time is spent in receiving oral than in sending oral. Since reading is at least as fast as writing, and since chemists read 1 1/2 times as often as they write (7.2 vs. 5.0) there must be several readers for each item written or else chemists read material written by non-chemists. With respect to oral communication this is reversed. More time is spent talking than listening. Since at least one listener is required for each talker, this must mean that chemists talk to non-chemists more often than they listen to them. Further analysis will explore this in more detail when we deal with the classification of the other people with whom the chemists communicate. This discussion is presented merely to indicate the kind of interpretations to which the data lead. The third, and perhaps least expected result shown by this table, is that written information exchange is used only one and a half times as much as oral. In this table, oral information exchange specifically excludes discussion, and is restricted to actual transfer of

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

FIGURE 4. Chemists’ allocation of communication time by direction of information flow.

information, rather than the give and take of a conversation. In view of the advantages of written communication (speed, clarity, the recipient’s choice of time, place, and speed) it seems that the large proportion of oral requires investigation. However, even this ratio (1 1/2:1) indicates that a given message now being transmitted orally would require less than half the total manhours, if it were to be transmitted in writing.

Interrelationships

To investigate the interrelationships among the various activities with respect to the time allocations, it was necessary to allow for the constraint imposed by the fact that the sum of the allocated time must equal 100%. Obviously as more time is spent in, say, scientific communication, less is left to be divided among the remaining activities. The definition of independence among activities used is: Activity Y is independent of activity X, if and only if, as X changes from the average, Y receives (or loses) time in proportion to the average Y. Mathematically this means that the expected value of Y is ,

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

FIGURE 5. Chemists’ allocation of communication time by person communicated with.

where =average value of Y, and =average value of X. It turns out that this is the same value to be expected under the multinomial hypothesis. See Appendix 4 for the mathematical statement of the multinomial and the derivation of the expected value under constraint.

With this expected value, interrelationships can be explored to see if, in fact, the actual values tend to exceed or to be less than the expected values. As an example, we will consider the hypothesis that the amount of time spent in Equipment use is dependent on the amount of time spent in Scientific communication. Figure 6 shows a plot of the raw data, the percent of time spent by each observed unit in each of these two categories. Here there is a definite indication that as Scientific communication increases, Equipment use decreases. But how much of this is due to the limit of 100% as a total?

Figure 7 shows the data replotted in terms that exclude this effect. The abscissa shows the deviation for each unit from the average Scientific communication (33.4%). The ordinate shows the deviation for each unit from the expected value of Equipment use; expected according to the definition of independence. This clearly shows that while Equipment use is dependent on

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

FIGURE 6. Relationship between Equipment use and Communication.

FIGURE 7. Unexplained Equipment use. Y, percent of time spent in Equipment use; E(Y), predicted percent of time spent in Equipment use. (See Appendix 4.)

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

Scientific communication, much of the apparent dependence was due to the 100% limit. As more time is spent in scientific communication, even less than would be expected is devoted to Equipment use. And as less time is spent in Scientific communication, Equipment use gets more than its share of the residual. Figure 8 shows that there is no effect on amount of Scientific communication by changes in equipment use.

FIGURE 8. Unexplained scientific communication. Y, percent of time spent in scientific communication; E(Y), predicted percent of time spent in Scientific communication. (See Appendix 4.)

In addition to the charts, a statistical analysis using chi square is shown in Table 4. This is the 2 by 2 cross-contingency table showing whether the number of points in each quadrant of the graph (as in Figs. 7 and 8) could have arisen due to chance, when in fact the control variable exerts no influence on deviations from the expected value of the dependent variable.

An additional series of tests were run to investigate the relation between these variables and some of the characteristics of the units observed. Appendix 5 contains a copy of a questionnaire sent to each cooperating company. At

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

TABLE 4 Chi-square tests on deviations from expected value of dependent activity

 

SC, Scientific communication; BC, Business communication; T and PA, Thinking or planning alone; ESU, equipment setup and maintenance; EU, Equipment use; DT, Data treatment; P and S, Personal and social, M, Miscellaneous.

In each square of the table: P=0.005, means the probability is 0.005 or less than the degree of dependence shown by the data is due to chance. If the probability was greater than 0.200, it was omitted; + or −, the relationship between the control and dependent variables was positive or negative; [8.088], the value of chi square for the indicated relationship, with one degree of freedom.

present only enough returns are in to permit partial analysis, but one of the most striking results is that none of the variables of time allocation is dependent on the number of chemists in the observed group. This group size varied from 5 to 111. Table 5 shows the relation between group size and each of the eight gross time allocations. The hypothesis that large units require more internal communication is scarcely borne out by these data. While some of the variables do show a slight dependence on size, it is not so large as was expected.

Additional analyses under way will investigate the relationships among the several components of communication, as well as with company characteristics such as basic vs. applied research, number and type of persons communicated with, and library facilities.

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

TABLE 5 Group size as it affects time allocation

Activity class

Largea companies above average time allocation

Smallb companies above average time allocation

Probabilityc of difference at least this large

Scientific communication

15

9

0.10

Business communication

10

10

1.00

Thinking and planning alone

5

11

0.10

Equipment setup

11

5

0.10

Equipment use

8

9

0.90

Data treatment

10

9

0.90

Personal and social

11

10

0.90

Miscellaneous

16

8

0.025

Out of area

15

10

0.20

a Companies with more than 20 “observed” chemists.

b Companies with less than 21 “observed” chemists. Total number of companies is 50.

c Computed from chi square 2×2 tables.

The major purpose of this phase of the analysis is to determine what will be the effect of changing the amount of communication. If the relative value of the various activities can be approximated (perhaps through experiments) then communication time can be increased at the expense of less valuable activities, or decreased in favor of more valuable ones. In any event, the value of communication (or any part of communication) must be meaningfully related to only those activities that vary non-randomly with it. When completed, a direction will have been established for the development of a measure of the value or worth of various kinds of written and oral communication.

Naturally, since the analysis is still in progress, the conclusions and interpretations are tentative; they will be confirmed, rejected, or modified on the basis of the further results, which will, we trust, be available in the very near future.

Intracompany relationships

Since there were 18 successive sets of observations made on each company, it is feasible to investigate the intracompany variability. One examination was made to see if there was any trend either in the averages or in the dispersion. This analysis was conducted on a sample of ten of the fifty companies, and a modified form will be extended to the rest. The results indicate that there was no time trend. This is evidence that our training sessions for the interviewers was adequate to bring them to an acceptable level where no further learning occurred during the data collection.

A second analysis of the intracompany data was performed to test the hypothesis that the 18 sets of data we obtained were, in effect, 18 random samples from the same population.

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

In order to run the test, the form 1 frequency data for each unit (see Appendix 1) were grouped into 6 activities within each of the 18 rounds: scientific communication, business communication, thinking or planning alone, equipment setup and data treatment,1 equipment use, personal and social. The other form 1 classifications, “none of these” and “out of area,” were not included in the test.

By classifying the data by activity and round number, a 6 by 18 matrix was formed. The row and column totals were used to compute expected values within each cell of the table. X2 was computed with 5×17=85 degrees of freedom.

Possible interpretation of the results are:

  1. Low values of X2 may be explained, as the result of task specialization. One chemist works mainly on equipment, another searches the literature, etc. This violates the “equal probability” assumption on which the test is based. It should be noted here that the hypothesis under test is whether all chemists within a unit are alike. This makes no statements about differences among companies.

  2. High values of X2 may be explained as the result of chemists performing certain activities in groups. Discussion, end of the month reports, and “personal and social” are obvious examples of this possibility. Here the assumption of independence (and hence the “random sample” assumption) is violated.

  3. Certain of the smaller units have low frequencies for some of their activities. The chi-square approximation often fails under such conditions and spurious values of X2 (usually too high) may result.

With X2 for 10 units calculated, all the values have been high, the lowest being 87.3. The results are tabulated in Table 6.

The effect of low frequencies is important with some of the smaller units,

TABLE 6 Chi-square results of intracompany analysis

Size of unit

Number of observations

Value of X2

Probability X2ν

9

147

104.40

<0.10

20

280

87.32

≈0.40

25

313

162.20

<<0.0001

31

429

98.61

<0.25

34

461

112.46

<0.025

35

508

118.13

=0.01

40

623

115.30

<0.025

63

814

135.14

<0.0001

99

1183

95.67

<0.25

111

1433

163.05

<<0.0001

1  

These two were grouped together because of their similarity of function and to increase the expected frequencies within each classification.

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

but the most important factor seems to be grouping. This is evidenced by the fact that large units show no tendency to have a lower X2 than small (low frequency) units, and also by the consistently high contribution of “personal and social” to X2, even when the expected values exceed 4 in every box.

Plans for additional analyses

In summary, two preliminary analyses have been conducted. The first was a search for interrelationships between classes of activity. In this analysis all observations for each group of chemists were combined into single estimates of the proportions of time spent in each activity. Then each group was treated as a single data point in looking for relationships between activities. The assumption was that, if there were no relationship between two activities, then, as the proportion of time spent in one of these activities moves away from the average proportion over all groups, the proportion of the residual time represented by the second activity will remain constant and equal to the average proportion of the residual for that activity over all the groups. As yet, the analysis has not isolated any strong relationships which would deny this assumption.

The second analysis tested the assumption that the chemists within a group can be viewed as a single population, each member of which distributes his time in accordance with the allocation typical of the group as a whole. If this assumption were true, the activity pattern during each of the 18 observation periods within a given group should be randomly distributed about the over-all activity pattern for the group. This assumption was not supported by the test. Instead there appeared to be a substantial “grouping” effect, that is, chemists seem to work at tasks in groups so that what one chemist does influences what another is doing.

A third analysis which is just starting will investigate whether there are significant differences in time distributions between groups. If there are, these results taken together with those from the second analysis would indicate that the environment within which a group of chemists work has a strong influence on how the group distributes its time. The environmental data being gathered should be useful in isolating those factors which affect the allocation and should indicate the nature of the relationship between a given factor and the group’s activity pattern. Knowledge of these factors and their influences on time allocation may provide means to manipulate the allocation in directions favorable to increased productivity.

A number of other potentially important analyses are contemplated. First, although a direct attempt to construct a measure of effectiveness based on

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

scientific productivity has been purposefully avoided, there are a number of broad-gauge effects of productivity which can be utilized in a search for useful hypotheses. For example, it seems reasonable that a chemical company which has experienced a significantly more rapid increase in sales volume over the past decade than could be explained by the normal expansion of the economy could very well have done so as a result of introducing successful new products. If these products were developed by the company’s research staff, then it must be considered relatively high on the scale of scientific productivity, in the applied sense, at least. Assuming a relatively stable protocol for scientific behavior within companies, comparisons between the activity distributions of chemist groups in highly productive companies with those in the other end of the scale may bring to light important effects. Further, adjusting for differences in work group environment may make these differences more apparent.

A second major analytical segment will be concerned with evaluating the relative efficiencies of various means of communication, and the effects of using more or less of the efficient media on the areas of activity from which scientific productivity must come. As an example of this type of analysis suppose that, in the area of scientific communication, the total amount of information received through written media is higher per unit of time spent in reading and writing than is the amount of information received orally per unit of time spent in talking and listening. Now suppose that the groups of chemists are arranged in descending order according to the proportion of each group’s total scientific communication time that is concerned with written communication. One possible result might be that these groups which spend more of their communication time in using written communication also spend a higher portion of non-scientific communication time in a combination of Thinking and planning alone, Equipment use, and Data analysis than do those whose Scientific communication is predominantly oral. If so, extension of this analysis into the content of the written communication should provide additional understanding of the relationship and, consequently, a higher probability of being able to modify the relationship favorably through changes in content.

In addition to using the content of communications in conjunction with the relative efficiency analysis described above, content analysis per se may provide broader comprehension of the scientific activity process. Relationships such as that between the portion of time spent in reading chemical abstracts versus the portion of time spent reading professional books and articles may indicate whether abstracts are used as a substitute for the professional literature on which they are based or whether they are used as guides to this literature. In other words, insight may be gained into the function of one form of scientific

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

communication with respect to the other forms. Furthermore, relationships can be sought between the form and content of a communication and the other aspects of the group’s activity pattern.

ACKNOWLEDGMENT

This project involved the efforts of many including Beverly Bond, Richard E. Deal, Philip Forsyth, Robert J.Frame, Michael Leyzorek, Joseph F.McCloskey, B.H.P.Rivett, George W.Summers, and Ronald W.Wolff, all from Case. The great contribution of Dr. Alberto Thompson and Mrs. Helen Brownson of the Office of Scientific Information, National Science Foundation, should also be cited. The study would not have been possible without the cooperation of the American Chemical Society and the many companies and universities and their personnel who permitted themselves to be observed.

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

APPENDIX 1

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

APPENDIX 2 Data collection instructions for project 957

  1. List of materials

    1. 25 copies of Casor form 957–1 and envelopes,

    2. A supply of Casor forms 957–2.

    3. A set of instructions.

  1. Observation procedure

    1. An observation is made on each selected chemist and consists of looking at what he is doing when you first come into his office, lab. or meet him in the hall on either your morning or afternoon round. Each observation is checked on form 957–1 and if it is “scientific communication,” then a form 957–2 is filled out.

    2. The list of “chemists” for this study is made up of full time employees who:

      1. are members (not student affiliates or student members) of A.C.S., or

      2. have a B.S. or higher degree in Chem. or Chem. E., and are currently working in chemistry (utilizing their technical training), or

      3. have a bachelor’s degree in some other science (math., biol., statistics, etc.), and have been working as chemists for at least one year, or

      4. have no degree in science, but have been working as chemists for three years.

        This definition is not intended to include technicians, machinists, laboratory assistants, etc. “Working as a chemist,” means that the level of skill and training required is that usually indicated by a B.S. degree.

    1. The chemists on the list are those who have their desks (or can usually be found) in the area that contains the traced chemist. This area is chosen to be feasible for observation, and contains no less than five chemists, but may contain as many as 100.

    2. An observational round starts at the time shown as “Scheduled start time” on the top of form 957–1. You go around the area and observe the chemists’ activities. It is easier if you carry a list of the chemists and cross off each one as you see him. If there is a chemist you don’t see, find out where he is, and if it is feasible, go and observe him, (he may be down the hall at another lab.). If he is out of the building or cannot be located, check “Out of area” on form 957–1.

    3. After each round, check the forms to see if you have filled out all the columns, and that you have a form 957–2 for each check under “scientific communication” on form 957–1, put the completed forms in an envelope, and relax (or go back to your regular work).

  1. Casor form 957–1. (See sample.)

    1. There is one form 957–1 for each round. There are two rounds a day, one in the morning and one in the afternoon. This form will contain a mark for each chemist on the list. Either he will be classified in one of the first eight rows (Scientific communication to “None of these”) or will be, “Out of area.”

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

The “total” number of checks must always be equal and be the number of chemists on your list. Every chemist must be accounted for.

  1. The heading of this form contains five items (see sample). The first, “No. _____” will be filled out when you get the forms. It indicates your organization’s code designation. The second item, “Round _____,” is to indicate which set of observations are recorded on the form. This is a two digit number. The second digit will always be “1” or “2.” “1” shows that the round was a morning round, “2” indicates an afternoon round. The first digit indicates the day. The first day on which observations are to be made will be day 1, the second day will be “2,” etc. Thus round “3 2” will be the afternoon round on the third day of data collection; round “8 1” will be the morning round on the eighth day, etc. “1 1” will be the first round and “9 2” will be the last.

  2. In the item “Observer _____” put your initials (or the initials of the person who makes a round in your place). This is so that if there are any questions they can be referred to the correct person.

  3. The next item is “Scheduled start time _____.” This is filled out and is either on the hour or the half-hour. These times were picked to give a representative picture of your working day, allowing for lunch. In the next blank, “Actual start time _____,” enter the actual time you start the round, to the nearest minute. We fully realize that you have your regular duties to attend to, and that there will be some rounds that will be started early or late. Please enter the actual time you start the round, even if it is quite different from the scheduled time. If you cannot make the round in the same half day as it is scheduled, write “Missed” in the “Actual start time _____” blank, and a make-up round will be added at the end of the regular nine-day period. e. The body of the form consists of ten rows, a space for counting marks for each row, and a “total” column. Do not use the column at the extreme right of the form (the one that starts “11, 12, 13”). In the large space, using single strokes with a diagonal for five to record the individual observations. For example, “///” is three, is seven, and is 12, etc.

  4. In the total column, record the number of marks in each row, and enter this number under “total.” This will make adding the grand total easier, and will facilitate further analyses.

  5. The categories can best be defined by examples.

    1. “Scientific communication” is meant to include all talking, listening, reading, writing, and discussion concerning technical or scientific matters related either to the job or to outside professional activities.

      Examples are: reading an article, handbook, memorandum, etc.; discussing work, project status, etc., talking on the phone to someone about a technical matter; writing an article, book, report, memorandum, etc.; editing or reviewing material for publication.

    2. “Non-scientific business communication” consists of all the necessary communication (business and professional) that is not of a scientific or technical nature. This category includes supervisory and administrative communications, letters, memos, and discussions relating to conducting the business activities or outside professional activities.

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

Examples are: discussing personnel, office space, price, delivery, vacation schedules, and staff meetings.

  1. “Thinking or planning alone” is the category for activities such as writing your own notes, working out a plan on paper, (or blackboard), or just sitting and thinking (about scientific matters). Some chemists work best in groups, some best alone. This category is not to be confused with “scientific communication,” which may well be used for activities done alone. If there is communication with someone else, by writing, reading, or phone, even though the chemist is alone, it is “scientific communication.” If he is alone and thinking or planning (not communicating with anyone else) then this category (3) applies.

    Examples are: designing an experiment, working out an idea, sketching an apparatus layout for his own use, and thinking thru a problem.

  2. “Equipment setup and maintenance” and

  3. “Equipment use” are two very important categories and will account for most of the observations. They include working with chemical and scientific apparatus. The two classifications distinguish between work that is preliminary (including calibration) and that which yields actual data. The setup and maintenance are necessary, but are to be coded separately from the use of the equipment.

    Examples are: calibrating a pH meter, bending tubing, recording readings on a data sheet, washing flasks, and titrating a solution.

  4. “Data treatment” is to be used for the analysis of numerical data such as calculating formulae, drawing graphs, making up tables, and charts. This category is for the statistical, numerical, mathematical treatment of observations, not for mere transcription or recording (see examples in 4, 5).

    Examples are: adding numbers, calculating averages, and drawing graphs.

  5. “Personal and social” is for the ordinary daily activities that chemists perform that are not actually part of their jobs. They do them because they are humans as well as chemists. This includes coffee breaks, talking about the weather, TV, politics, or the wife and children. Also included are personal or social communications such as writing personal letters, checks, calling friends or family. Be sure there is no attempt either to minimize or to exaggerate this category. There is not the slightest implication that personal and social activities are undesirable. They are part of the everyday business of living and getting along with others.

  6. “None of these” is a miscellaneous category for cases that do not fall into any of the others. It includes walking in the halls, waiting, and straightening up a desk. In every case where it is used, make a note on the back of the form explaining what the observation was. Use this category only if you cannot fit the observation into any of the other-categories.

  7. “Out of area” is for those cases where you cannot locate the chemist or you know that he is out of the building, home, sick, on vacation, on a business trip, etc. If you can categorize his activity then do not list him here, but put the check in the appropriate box. If he left the office for a doctor’s appointment, for example, check “Personal and social.” If he

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

went to the library check “Scientific communication,” but on the 957–2 just write in “At Library, not observed.” You may have to ask to find out where he went.

  1. For “None of these” write details on back of form.

  1. Casor form 957–2.

    One copy of this form must be filled out for each mark under “Scientific Communication” on form 957–1.

    1. The heading “No: _____, Round: _____” should be filled out to agree with the same entries on form 957–1. These forms will be kept together at the end of each round. There will be one 957–1, and as many 957–2’s as there are checks under “Scientific communication” on 957–1.

      When filling out a 957–2, there must be one entry in the first column, and two entries in the second column, one under “Number and type of people involved in the scientific communication” and one under, “Form of the scientific communication.”

    2. In the first column an entry is made by circling the number at the right of the entry, after the word “from” or “to” or “with.” The purpose of the word is to remind you that the entries in the second column go on the left or right, depending on whether they refer to a “from” or a “to, with.” In this second column, an entry will be a number written in by you in the appropriate box. This number tells how many people the observed chemist is communicating with. If he is discussing something with a group, then you write in the number of people (do not include the observed chemist in this count). On every form 957–2 there will be at least one entry in this section. There may be several, since the other people being communicated with may be distributed among the categories; for example, some biologists and some chemists might be discussing something with the observed chemist. If the observed chemist is reading, then the number and classification of “Number and type of people …” would refer to the author or authors. It is extremely important that these numbers be accurate. The only time numbers will not be used is when material is being sent to a large and undetermined audience, as in the case of an article being written for a journal or a speech being prepared for a professional meeting. In that case, use the letter Y and put it in the appropriate box or boxes. If you have no idea as to the composition of the audience, use “none of these” and describe the situation on the back of the form.

      In the last section (bottom) of column two, an entry is made by circling a number (or X or Y).

      In the two places (column one and bottom of column two) where there are numbers printed in, an entry consists of circling the number. In the other case (top half of column two) where there is no number, write one in.

    3. Column 1. Description of scientific communication

      1. Hearing question refers to occasions where the observed chemist is listening to an oral request or question. The actual words do not have to be in question form. It may be a suggestion, or request. The important aspect for this classification is that the answer or reply would normally contain scientific information. This category is not to be used for casual questions

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

that occur in discussion like, “Don’t you think so, Joe?” or, “Well, then, is it agreed that well get together at 10:00 on Tuesday?”

This classification should be used for oral requests for technical information, for example: “Joe, would you know what journal had that article by Klumpenmeyer on semi-captive radicals?” This category is also used for questions about apparatus, calibration, etc. If the question comes up in a discussion, do not use this category, use one of the two discussion categories. This category is for questions or requests for information that are the purpose of whole communication. Note that the entries in column 2 will be in the “from” column and that “(1) oral” will be recorded at the bottom of column 2.

  1. Reading question is the same as hearing question, except that it refers to reading a question or a request for information.

  2. Reading for use. This classification is for the reading of memoranda, articles, books, etc. when the purpose of the reading is to find some specific information of direct use on a current task. Examples include looking up values in a handbook and checking a formula in a text or reference book.

  3. Reading for general information. This category is for all the scientific reading that is done for background, for its general value, because you get the journal anyway, or just because it is interesting. The reading must be scientific or a 957–2 would not be filled out at all. (Popular Mechanics is not scientific literature.)

  4. Hearing information is used for two kinds of situations. One is in the answer or reply to a question. Remember this refers only to those situations where the main purpose of the interchange was the answering of a particular question or request, not those questions arising during a discussion. The second situation is in listening to a lecture, briefing session or talk where there is little chance to break in with questions or remarks. This use of “hearing information” is used to denote a one way communication system. If the exchange is two way, then the category to be used is “discussion.”

    Examples of “hearing information” are: Listening to the answer to a technical question, being briefed on a new technical development by a visiting expert, and listening to a report of a professional meeting by someone who attended.

  5. Working out material is used for those situations where the observed chemist is trying to understand something he has read or heard. He may be using a blackboard or pad to see if a certain reaction will follow as Hemplemeyer says in his article. In the later parts of the recording form the data refer to the source of the material he is working out.

  6. Editing material received is for those cases in which the observed chemist is proofreading, editing, refereeing, or otherwise working with material for publication (either for journals or books, or for publication within the company). This may be something he has written himself or something that someone else wrote, but it is now being readied in final form,

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

not still being written. In the last section of 957–2, use the category that refers to the ultimate form. If a chemist is editing material to be submitted as an article to A.C.S., this would be called “article” even though it is as yet unpublished. If he is reworking a speech for future delivery it is “oral” even though the speech may eventually be published.

All the above categories are “from” and must have entries under “From” in the second column on the form.

  1. Writing information is used for all written scientific communication including letters, memoranda, reports, articles or books. It refers to outgoing information only and not to written requests for information.

  2. Telling information is the same as “writing information” except it refers to oral methods rather than written. If one person is “hearing information” (as discussed above) then someone else must be “telling information.”

  3. Writing question refers to writing a request for information.

  4. Asking question is the same as “writing question” but is oral, not written. It is the other half of “hearing question.”

    The previous four categories are all “to,” and must have their entries in column two on the right-hand side, “To, With.”

  5. General discussion is that part of scientific communication that goes on in groups. It is not “one way” (like hearing or telling information). The discussion may be about some highly specific point, but it must be a real discussion, not a lecture. Note that this category is “with,” and requires the number and kind of co-discussants.

  6. Discussion about a received communication is similar to “working out material” but refers to a “bull session” where the participants are trying to understand or put to use something they were told or read. This category is “from” and “with.” It requires entries on both sides of column two.

  7. Reading for retransmittal refers to the case of an observed chemist reading an article or book, etc., to report on it (either orally or in writing). In this case as in the previous one, there will be entries on both sides of column two, “From” and “To.”

  8. None of these is for any difficult case that doesn’t fit in. Be sure to write in the details on the back of the form.

  1. Column 2. Number and type of people involved in the scientific communication.

    In column two of form 957–2, an entry is a number. Write in a number in the box, using either the left or right set according to the “from” or “to” in the entry you circled in column one.

    1. Mathematicians, statisticians include all the mathematical sciences, experimental designers, and geometers.

    2. Physical scientists include physicists, metallurgists, astronomers, optical scientists, meteorologists, and electronics men.

    3. Chemists, same unit refers to chemists who are in the same project group, or in the same administrative unit (where this is defined as narrowly as possible). The intent here is to include men working on the same project, or a group of closely related projects.

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
  1. Chemists, different unit refers to other men who are chemists, but who are working on different projects or different types of problems. If a biochemist working on new insecticides is talking with a physical chemist working on artificial crystals, they are in different units. (For “unit” you can read “branch of chemistry.”)

  2. Biologists, medical men include what are often called the life scientists. These are biologists, pharmacologists, zoologists, botanists, medical men, etc.

  3. Behavioral scientists are the group containing psychologists, economists, political scientists, semanticists, etc.

    If you are in doubt, write in the name of the science and then the number of such scientists involved.

  4. Secretaries, technicians, etc., include students, lab. assistants, and machinists. These are the people that directly help the chemist get his work done.

  5. Other company personnel are your company’s sales men, executives, administrators, staff assistants, etc.

  6. None of these includes customers, visitors, outsiders of all kinds. Be sure to write in an explanation on the back of the form,

  1. Form of the scientific communication.

    The last section of form 957–2 deals with the nature of the communication. Again check in the appropriate side (“From” or “To, With”). In the case of “Reading for Retransmittal” there will be a check on each side, one for the nature of the reading, one for the form of the retransmittal. “Unpublished written” includes letters, memoranda, and reports for company or client use. “Published” (book, article, abstract or review) means that the material will be generally available to chemists (often with difficulty—as a Ph.D. thesis).

    In every case where anything but “oral,” or, “Unpublished written” is checked a reference must be put on the back of the form showing the title, journal, issue, or date, and page number. This is extremely important. Get the complete reference, so if we go to a library we can find it.

    In every case on the form where “None of these” is checked, write in the details on the back of the form.

  1. Security

    Since no names or code numbers are associated with individual chemists, there is no possibility (and certainly no intention) of identifying any chemist.

    A copy of the final report will be given to your organization, but it will not report data for any individual or company. No one in your organization except you will see the recording forms. Thus the integrity and individuality of your chemists is guaranteed.

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

APPENDIX 3

  1. The Universe to be sampled (target universe) is the set of Chemist-Minutes constrained by the following restrictions:

    1. Continental U.S.

    2. 5-day week, normal work hours/day.

    3. Work in one of the 150 metropolitan areas (1950 census definition) listed in Table 73, Detailed Characteristics, U.S. Census of Population.

    4. Work for a unit (smallest administrative unit in a company) employing 5 or more chemists (members of A.C.S. as of June 1, 1957). [Note. Restriction 3 limits the sampled universe to about 80% of the target universe. Restriction 4 limits the sampled universe to about 85% of the target. If they are uncorrelated, then, the restriction is to about 68%. However, there is probably some positive correlation (larger companies are more likely to be in metropolitan areas) so 68% is a minimum estimate.]

    5. The time period—July to October 1957.

  1. The procedure used to generate a sample of Chemist-Minutes is as follows:

    1. The hundred fifty metro areas were ordered from most chemists to least chemists. A systematic stratified sample of 10 areas was drawn which yielded (1) New York, (2) New York, (3) Chicago, (4) Philadelphia, (5) San Francisco, (6) Wilmington, (7) Buffalo, (8) Albany, (9) Kalamazoo, (10) Trenton. This sample was based on the 1950 census data (Table 73, Detailed Characteristics).

    2. In each metro area, the names of 50 chemists were drawn at random (systematic random sample—no stratification). Data used was the current A. C. S. roster.

    3. Each name drawn was checked to get number of chemists employed by the firm (same general address—to treat individually, multi-plant firms) and occupational classification of chemist. If firm employed ≥5 and occupation was in (see Appendix A for list of occupational classification), then firm (unit) was included in the sample. This operation was continued until approximately 100 chemists (totaling the number of “in” chemists employed by each unit) were obtained.

    4. This generated a sample of approximately 60 firms and 100 chemists.

    5. Each chemist in each unit was observed at two random times (one before lunch, one after) for nine consecutive work days. These observations (approximately 18,000) form the basis of the further analysis.

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

APPENDIX 3a

The sampling procedure yielded a sample with the following characteristics: number of companies, 42; number of observed units, 50; number of chemists, 1305.

Breakdown by Metropolitan area:

Metro

Companies

Units

Chemists

New York

11

15

257

Chicago

6

6

159

Philadelphia

2

2

114

San Francisco

6

6

147

Wilmington

2

2

160

Buffalo

4

4

89

Albany

6

7

179

Trenton

4

5

96

Kalamazoo

1

3

104

The sample was designed to obtain approximately 100 chemists in each area (200 in New York since it was considered two areas). The size of the units varied from 5–111 chemists with most of them lying within a range of 10–40.

APPENDIX 4 The use of the multinomial assumption

Since we have a universe of chemists one assumption we may make about the underlying structure is that each and every chemist has the same probability for any particular activity as any other chemist, and that these probabilities do not change with time.

Then, if we were to observe a group of chemists, we could predict, on the average, the kinds of observations we should get.

Thus, if there are k different, mutually exclusive and exhaustive activities, and associated with each one is a probability (P1, P2, P3, · · ·, Pi, · · ·, Pk) such that

then if we observe n chemists at random, the probability that we will get exactly x1 observations of the first activity, x2 of the second, etc., and xk of the kth activity is given by:

This is the multinomial distribution.

Our concern in this situation, however, is to determine the distribution of x1 given x2, i.e., Pr{x1|x2}. This can be done simply by applying conditional probability to the multinomial for x1, x2, and x3.

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

Letting M(x1, x2)=the multinomial for x1, x2, we have

which equals

which reduces to the expected value of x1|x2:

This can also be expressed as

or in percent terms

which is the computational form used in the body of the report.

The variance of

or, the standard deviation, in percent terms is:

These two forms, the percent forms for expected value and standard deviation, are the ones utilized for the computations in the report.

APPENDIX 5 Scientific Communication Study—Questionnaire

  1. Would you please classify the major activity of the “observed chemists” as:

    ☐ basic research

    ☐ applied research

  2. Would you also classify the nature of the research conducted by them as being primarily either

    ☐ product research, or

    ☐ process research

  3. Please check any of the following categories that describe the literature facilities available to the “observed chemists.” You may check as many or as few of the categories as apply.

    ☐ Current journals available to the men at their desks. (Personal subscriptions or routed copies.)

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×

☐ Library facilities in the same building with the literature available for withdrawal.

☐ Main company library not in the same building, but literature can be withdrawn on request.

☐ Public library facilities available with a procedure for company use.

  1. What is the average annual cost of the physical plant, exclusive of the laboratory equipment and materials, used by the “observed chemists?” Please include the amortization of the capital goods, any imputed rent, and operating and maintenance costs. $______/yr.

    What is the average annual cost of the laboratory equipment and supplies used by the “observed chemists?” Include any amortization of capital equipment in this figure. $______/yr.

    [If the “observed chemists” do not have separate facilities, apportion the costs on a manpower basis.]

  2. This question refers to a group of people which includes the chemists who were observed for two weeks earlier this year. We want to include in this group most of the people who work and communicate with the “observed chemists.” This will usually be all of the men in the laboratory, or all of the men in same building, but it can also include men from other units of the company, if they are in frequent contact with the “observed chemists.”

Classification by specialty

Number in group

 

Average annual salary

Mathematicians and statisticians

________

 

$_______

Physical scientists, including physicists and metallurgists

_______

Chemists, including both “observed chemists” and all others

_______

Chemical engineers

_______

Other engineers

_______

Biologists, medical men, life scientists

_______

Behavioral scientists, including psychologists and economists

_______

Technicians, laboratory assistants

_______

 

$_______

Secretaries, clerical staff

_______

 

$_______

Administration personnel

_______

 

$_______

Total in group

_______

 

 

  1. There were “observed chemists” in your group, and your observer has a list of their names. Please indicate here their average annual salary: $_______.

  2. Please write here any comments you wish to make about this questionnaire:

    ____________________________________________

    ____________________________________________

    ____________________________________________

    ____________________________________________

    ____________________________________________

Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 97
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 98
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 99
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 100
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 101
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 102
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 103
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 104
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 105
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 106
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 107
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 108
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 109
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 110
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 111
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 112
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 113
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 114
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 115
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 116
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 117
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 118
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 119
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 120
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 121
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 122
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 123
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 124
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 125
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 126
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 127
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 128
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 129
Suggested Citation:"An Operations Research Study of the Dissemination of Scientific Information." National Research Council. 1959. Proceedings of the International Conference on Scientific Information: Two Volumes. Washington, DC: The National Academies Press. doi: 10.17226/10866.
×
Page 130
Next: Information and Literature Use in a Research and Development Organization »
Proceedings of the International Conference on Scientific Information: Two Volumes Get This Book
×
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

The launch of Sputnik caused a flurry of governmental activity in science information. The 1958 International Conference on Scientific Information (ICSI) was held in Washington from Nov.16-21 1958 and sponsored by NSF, NAS, and American Documentation Institute, the predecessor to the American Society for Information Science. In 1959, 20,000 copies of the two volume proceedings were published by NAS and included 75 papers (1600 pages) by dozens of pioneers from seven areas such as:

  • Literature and reference needs of scientists
  • Function and effectiveness of A & I services
  • Effectiveness of Monographs, Compendia, and Specialized Centers
  • Organization of information for storage and search: comparative characteristics of existing systems
  • Organization of information for storage and retrospective search: intellectual problems and equipment considerations
  • Organization of information for storage and retrospective search: possibility for a general theory
  • Responsibilities of Government, Societies, Universities, and industry for improved information services and research.

It is now an out of print classic in the field of science information studies.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    Switch between the Original Pages, where you can read the report as it appeared in print, and Text Pages for the web version, where you can highlight and search the text.

    « Back Next »
  6. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  7. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  8. ×

    View our suggested citation for this chapter.

    « Back Next »
  9. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!