ABSTRACT THE SAWMILL MANAGER: HIS NATURE AND HIS TIME BY Antone Cornelis Van Vliet An investigation was made of 11 managers of small sawmills in Michigan concerning their nature and their use of time. The time these managers devoted to decision making, communication, and other characteristic activities was studied. The study is divided into three parts. Part I involved the choice of work sampling as a technique used to collect data. A self observational variation of work sampling was made possible by the invention of a new Random Interval Timer. This compact device emitted random signals at a rate that could be pre-set by the researcher. The Poisson process was chosen as best fulfilling the technical requirements. The 11 managers carried their instru- ments an“ average of 25 days, spanning four months. They produced 2215 total observations in 11 activities, each recorded by time of occurrence, type of activity, and duration of event. Part II involved in-depth interviews with each manager to see if different types of managers existed. Generally, these men could be f! “7/" ”'1“: {/5/ Antone CornelisVan Vliet best described as exhibiting an entrepreneurial behavior. This dis- covery led to using a method of separating the managers into the Craftsman-Entrepreneur (C-E) or Opportunistic-Entrepreneur (O-E) classifications. If these classifications were valid, then a series of hypotheses could be tested using time-use data. The findings were discussed in three forms: 1) profile of the entire group, 2) three High O-E vs. the three High C-E managers, and 3) three High O-E vs. all the C-E managers. Hypothesis: The Opportunistic-Entrepreneur spends more time in the decision making process than the Craftsman-Entrepreneur. Result: The three High O-E's spent 114% more time in the decision making process than the three High C-E's. This proved significant at the 5% and 1% levels only for the pure types, not for the O~E's vs. all C-E's. Hypothesis: The Opportunistic-Entrepreneur spends more time in the communication process than the Craftsman-Entrepreneur. Result: The three High O-E's were involved in 2.4 times as many communication interactions as the three High C-E's. Hypothesis: The Craftsman-Entrepreneur engages in more manual labor in his firm than the Opportunistic-Entrepreneur. Result: The three High C-E's spent 309% more time doing manual labor than the three High O-E's. Hypothesis: The Opportunistic-Entrepreneur participates in more training programs than the Craftsman-Entrepreneur. Antone Cornelis Van Vliet Result: Out of a total of 2215 observations, only two events involved training. An activity profile of the 11 sawmill managers revealed that they spent almost one third of their time doing routine work. Manual work accounted for 10.88% and the decision making process (four segments) occupied 27. 47% of their time. Duration time for each activity was also calculated. Part III involved looking at the communication networks and organizational structure. The entire group of managers were com- municating in 35.45% of the total observations. In these instances. the manager was the sender over 80% of the time. The object or source of the communication contacts were listed by groups, such as employees, foreman, and equipment suppliers. Six organization charts are shown to understand how the sawmill manager views his firm. Conclusion. The self observation form of work sampling, coupled with the use of the new Random Inte rval Timer proved an effective method for collecting data. Their use could save large sums of money in studies that require accurately measuring a variety of tasks. The activity that was lacking--training--is a cause for concern. The possibility of discovering the right ”wave-bands” for reaching the managers involves the communication networks. The entrepreneur is a vital part of the free enterprise system. Antone Cornelis Van Vliet He exhibits innovation, courageous risk taking and energy. These values should not be lost for the sake of the large firm's efficiency. APPROVED: Major Professor Department Chairman THE SAWMILL MANAGER: HIS NATURE AND HIS TIME BY Antone Cornelis Van Vliet A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Forestry 1970 DEDICATION To Louise, Dan, Sue, Mary, Bill and Tom ii ACKNOWLEDGMENTS Many people helped in the formulation of this thesis with advice and consultation. I am grateful to Dr. Aubrey Wylie and Dr. Alex Panshin for encouraging me to attend Michigan State University and study in the Forest Products Department. Their friendship and counsel along with other members of the department, Professors Behr, Harold, Huber, Lloyd, and Sliker helped immeasurably in those moments of doubt. A special thanks goes to Dr. Otto Suchsland, major professor; to Dr. Miles Martin for introducing me to new methods of conducting scientific investigations, and his friends at Merck. Sharp and Dohme for the early Random Alarm Mechanisms; to Dr. Carl Frost for his personal interest; and to Dr. Darab Unwalla for his helpful advice. I am grateful and indebted to Joe Bauer, Russ Bradish, David Dix, Jim Devereaux, Ralph Devereaux, John Hawkins, Bob Johnson, George Reneand, Roy Richardson, Marvin Schneider, Cecil Szepanski, and Ken Thomas, without whose participation there would have been no study about sawmill managers. I am also grateful to the McIntire-Stennis Cooperative Forestry Research program, without whose funds this research could not have iii been conducted in Michigan. I am indebted to Mrs. Carol Thomas and Mrs. Amy Kaun. application programmers of the Michigan State University Computer Center, and to Dr. Norbert Hartmann and Dr. Scott Overton, Oregon State University statisticians for their helpful assistance. For their consideration, time and encouragement, a special thanks to Dr. Leon Garoian,and Joe Cox, Associate Director of the O. S. U. Extension Service. I thank my wife, Louise, for her typing and editing; MaryBelle Gros Jacques and Dr. Harold Laursen for proofreading, and Mrs. Clover Redfern for her final typing. iv TABLE OF CONTENTS Chapte r Page I. INTRODUCTION The Purpose of the Study The Forest Products Industry in General small Sawmills Some background information Loss of mills Managerial Time-Use PART I. ME THODO LOG Y II. THE SAMPLING PLANS Selection of Managers and Their Mills Work Sampling as a Technique What is work sampling and how can it be used? Advantages and limitations of work sampling An alternative: self observation system Pilot studies General Work Sampling Procedures and Cove rage Data Card design Proposed vs. actual pattern of cove rage Accuracy of Work Sampling Measurement III. THE RANDOM INTERVAL TIMER History of the Instrument General Requirements for Design of the Instrument Technical Requirements for Design of the Instrument The theoretical model The new Random Interval Timer Testing the Instrument PART II. TYPES OF MANAGERS AND THEIR TIME-USE IV. TYPING THE MANAGER Introduction The Entrepreneurial Manager The Craftsman-Entrepreneur and the Opportunistic- Entrepreneur Gathering the Information 11 ll 14 14 15 l7 19 20 21 23 29 33 33 34 35 36 39 46 59 59 60 62 63 Chapte r Oral interview Written interview Analyzing the Interview Material Formal education Work experience Social involvement Communication ability Delegation of authority Hiring of employees Capital sources Sales promotion Competitive strategies Future plans Amount of planning prior to initiation of the company Plotting the Sub-Types V. RESULTS OF TIME-USE STUDY ”Zeit ist Geld'I Hypotheses Tested in This Study Percent Time-Use All eleven managers High Opportunistic-Entrepreneur vs. High Craftsman-Entrepreneur Ave rage time allocated to activities High Opportunistic-Entrepreneurs vs. all Craftsman-Entrepreneurs Statistics used in comparisons The Results and the Hypotheses Ave rage Duration of Each Event Comparison With Existing Studies PART III. MORE DIFFERENCES: COMMUNICATION. ORGANIZATION VI. TYPE OF MANAGER: HIS COMMUNICATION AND ORGANIZATION PATTERNS Integration of Communication and Organization Theory The traditional approach Evolution of modern theory Finally some recognition Communication and the Sawmill Manager Total time spent in communication Direction of flow vi 72 73 75 75 77 78 78 81 85 88 88 93 96 101 104 104 104 105 105 106 108 110 Chapte r Favorite information sources-—work sampling A11 managers High O-E vs. High C-E Fri"orite information sources--interview data Organization and the Sawmill Manager The three High Craftsman-Entrepreneurs The three High Opportunistic-Entrepreneurs A cents-ible comparison VII. CONCLUSIONS Technical Contributions of the Study The Behavioral Contributions of the Study The three High Opportunistic-Entrepreneurs The three High Craftsman-Entrepreneurs Training--The Missing Activity The sawmill manager and training A possibility: the "change agent" Survival : Success or Success : Survival The Need for the Entrepreneur Future Research LIST OF REFERENCES APPENDICES Appendix A--List of Sawmill Managers in Study Appendix B--Sawmill Manager Information and Instruction Sheet Appendix C--Sawmi11 Manager Interview Form Appendix D--Cumulative Distribution Plots of Signals for Managers (Actual vs. Theoretical) Appendix E--Manage rial Activity Tables Appendix F--Tab1e of Student's t Test Values Appendix G--Tabu1ation of Data from Prototype Instrument Serial No. 81 Page 115 115 115 116 118 119 119 122 124 124 126 127 127 129 129 130 132 133 134 135 141 141 143 145 148 158 164 165 , w b Table 10. 11. 12. 13. 14. 15. LIST OF TABLES Page U. S. sawmill size classification. 6 Proposed plan and actual coverage of managers by months and weeks. 24 Total distribution of observations (N) by months, weeks and days. 25 Accuracy table for the daytime activities of 11 sawmill 31 managers. The mathematical comparison of the binomial and Poisson probability distributions. Distribution of signals per time period, actual vs. theoretical for SN 81. The calculated percentage of occurrences for time intervals (between signals) in the Poisson process-- actual vs. theoretical for test instrument SN 81. Calculations of F in minutes. Sub—type analysis of 11 sawmill managers. Comparison of ratio-delay method vs. pooled ave rages for each activity. Tests for significance by activity. Percent of executive time spent by activity—-7 studies. Percent time spent in communication. Tabular recording of communication flow. Favorite information sources listed by managers. viii 45 50 52 54 66 91 93 102 109 111 117 Appendix Table 3 A1. El. E2. E3. E4. F1. G1. List of sawmill managers in study. Number of daytime observations per activity per 'nanager. Number of nighttime observations per activity per manager. Percent of daytime observations per activity per manager. Average duration time (in minutes) spent per activity per manager--daytime observations. Table of Student's t test values. Tabulation of data from prototype instrument Serial No. 81. ix Page 141 158 160 162 163 164 165 LIST OF FIGURES Figure 1. 2. 10. 11. 12. 13. 14. 15. Location of 11 sawmills in Michigan study. Example of Data Card. Time signals recorded during each hour of the day (11 managers). Examples of the distribution of signals per hour of three managers. Block diagram of new Random Interval Timer and. wave-forms. Photograph of the Random Interval Timer. Poisson, binomial probability distributions of the Random Interval Timer. Three characteristic curves of the Poisson process. Distribution of signals per time period--actua1 vs. theoretical for SN 81. Distribution of signal intervals for the Random Interval Timer--actual vs. theoretical. Expected vs. actual number of signals of SN 81. Cumulative distribution plot of signals for manager No. 1 (actual vs. theoretical). Distribution of entrepreneurial types. Activity profile of all 11 sawmill managers. Percent of time spent in 11 major activities by High O-E's vs. High C-E's. 27 28 40 41 44 47 49 53 55 57‘ 74 79 82 Figure 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. -Actual number of observations recorded by High O-E's vs. High C-E's. fotal number of minutes per day devoted to key activities by three High O-E managers. Total number of minutes per day devoted to key activities by three High C-E managers. Percent of time spent in 11 major activities by O-E vs. all C-E managers. Average duration of each activity in minutes, comparing High O-E's vs. High C-E's and all C-E's. Average duration of each activity in minutes for 10 managers. Communication flow for all managers--shown as a percentage of 684 total observations. Communication flow for three High O-E type managers—- shown as a percentage of 248 total observations. Communication flow for three High C-E type managers-- shown as a percentage of 103 total observations. Organization charts--3 High C-E type managers. Organization charts--3 High O-E type managers. Sales comparison of selected firms (mills 2, 8, 10 vs. 4’ 6, 7)- Appendix Figure 3 D1. D2. D3. Cumulative distribution plot of signals for manager No. 2 (actual vs. theoretical). Cumulative distribution plot of signals for manager No. 3 (actual vs. theoretical). Cumulative distribution plot of signals for manager No. 4 (actual vs. theoretical). xi 86 87 89 98 99 112 113 114 120 121 123 148 149 150 Figure D4. D6. D7. D8. D9. D10. Cumulative distribution plot of signals for manager No. 5 (actual vs. theoretical). Cumulative distribution plot of signals for manager No. 6 (actual vs. theoretical). Cumulative distribution plot of signals for manager No. 7 (actual vs. theoretical). Cumulative distribution plot of signals for manager No. 8 (actual vs. theoretical). Cumulative distribution plot of signals for manager No. 9 (actual vs. theoretical). Cumulative distribution plot of signals for manager No. 10 (actual vs. theoretical). Cumulative distribution plot of signals for manager No. 11 (actual vs. theoretical). 152 153 154 155 156 157 I. INTRODUCTION The Purpose of the Study This is a study of the managers of 11 small hardwood sawmills in Michigan. 1 The purpose is to investigate how these managers spend their time, how they communicate on the job, and how they view their organizations. Many people familiar with the lumber industry have long sus- pected that there may be characteristic differences among sawmill operator's approach to management. These differences, if they exist, could provide a valuable base of information about this segment of the Forest Products industry. The process of narrowing the area of concern to this representa- tive managerial group is in response to what has been taking place in the small sawmill business across the country. Is the high mortality rate among small sawmills an inescapable consequence of the times or a condition that can be halted through a better understanding of these managers ? The national picture is mirrored in many ways by the Michigan sample. This research was carried out in Michigan with funds made available by the McIntire-Stennis Cooperative Forestry Research Program. 1 2 This chapter will briefly acquaint the reader with the Forest Products industry in general, and small sawmills in particular. The Forest Products Industry in General The Forest Products industry today is involved in a crucial per- iod in its history--a period that will see many firms fail to survive. The coroner's report will probably give numerous reasons for the deaths of those who finally close their doors or are absorbed by other companies. It may list among the causes the difficulty of acquiring raw materials, the growing obsolescence of men's skills and his machines, the keen intra-industry competition between the various wood products, the growing competition from outside industries (non- wood substitutes like plastics), the rising cost of labor, and the scarcity of capital or capital sources. Most managers agree that running today's firm is a complex and bewildering task. In a short span of time many new kinds of de- mands have been placed upon the modern executive. He must be more efficient, make decisions faster, possess or have access to greater knowledge, communicate with clarity and ease, and be willing to initiate and accept change--all at the same time. The Forest Products industry possesses some interesting characteristics different from many other major industries. Its common base is a renewable resource. timber, and the spectrum of industrial integration is unusually wide—-from rough lumber to engineered laminates; from chips to high grade paper; from thin veneer to attractive furniture. Segments of the industry, like lumber and plywood manufactur- ing, are tied closely to the monetary restrictions or seasonal fluctua- tions of the building industry. Lumber has been called one of the last true ”market adminis- tered'' price commodities in the United States economy. Some operators are extremely proud to be identified as one of these "last ' In economic terms, Mead (30) refers to bastions of free enterprise.’ the lumber industry as an excellent example of a real world approxi- mation to the economic model of pure competition. Technological innovation in general has been slow when com- pared to such rapid growth industries as electronics, chemicals or plastics. This fact alone may have lulled some operators into think- ing that "change' was an unnecessary ingredient for survival or growth . l . Small Sawmills Not only is there a wide diversity between segments of the For- est Products industry, but also within each basic commodity group. Sawmills, as a group, are no exception. There are so many sizes and For the purposes of this study, the term 'small' will indicate mills producing up to 5 million board feet annually; terminology indi- cating size of mills is not standardized nationwide. 4 kinds that it was necessary to concentrate upon one sub—group. Small sawmills have been singled out for this study because: 1. This group for many years has constituted a majority of the mills in the United States. 2. The number of small mills has been sharply decreasing over the years. In 1964 they contributed less than 30% of the national output of lumber. Some background information From the first sawmills established at Jamestown and later in New England, there has been a steady migration of mills following the westward movement and the development of the United States. On the heels of the pioneers, the mills came with the waves of goods and ser- vices that built the towns and made life bearable. These mills were established near the source of accessible virgin timber. Today the largest concentrations of mills are found in the west and south, close to the remaining resources. Lumber is a primary product; this is simply another way of saying that it is recognizable in its final form after it has gone through the mill. The resulting basic product has not changed substantially since the first mills were established. It was basically a good pro- duct with a variety of uses, and the sawmiller has been content to allow other craftsmen to capitalize upon its versatility. The majority of studies about the lumber industry have been done by economists. This is not surprising as the economist is well equipped to view an industry in terms of resource allocation, price systems and the market place. Economic studies provide some useful statistics and theories concerning the behavior of firms comprising an industry. In a general industry description, Keppler (20) states that in the period from 1929 to 1954 sawmills more than doubled in number in the United States while total lumber production was showing a slight de- cline. Acknowledging the difficulty of accurately determining mill numbers, he disclosed the existence of over 50, 000 sawmills in the United States according to the 1947 census of manufacturers. Since the early fifties production has remained fairly constant nationwide (between 32 and 37 billion board feet annually), but the number of pro- ducing mills has declined sharply. The large mills and the efficient medium-sized mills have increased in number and production. At the same time some of the small mills have moved into the next size class by improving their efficiency through modernization and expan- sion. The most current figures available concerning size classifica- tion of mills are for 1969 as reported in the May 29, 1970 issue of Forest Industries (22). While these figures are too incomplete to be useful here (only 64% of the production was accounted for), the trend is definitely continuing with a sharp decline for small mills. The 6 following table illustrates this decline from 1947 to 1964. 1 Table l. U. S. sawmill size classification. Mill Class 1 2 Annual Cut Mill Data 1947 1964 LARGE: Number of Mills 165 287 25 million board % of total number 0. 3 2. 5 feet and over % of total production 23. 0 42. 6 MEDIUM: Number of mills 939 1, 008 5 to 25 million % of total number 1.8 8. 6 board feet % of total production 25. 9 29. 0 SMALL: Number of mills 52, 005 10, 386 under 5 million % of total number 97. 9 88. 9 board feet ‘70 of total production 51. l 28. 4 Total number of mills 53, 099 11, 681 1Keppler (20). 2U. S. Bureau of Census (41). Loss of mills The loss of mills is a concern to many students of the industry. Why are so many of these mills now silent? Should we be concerned? Or should we simply tally these lost businesses to changing times? Many observers claim that the loss of small mills is inevitable and predict a continuance until the manufacture of lumber is in the hands of relatively few manufactures, an oligopoly. The economic ”facts of life" faced by small firms seems 1964 represents the most recent complete data available at the time of this writing. formidable. In addition to what has been mentioned the small mill manager, striving to survive, must deal with: 1. The growing competition for raw material at reasonable prices. This does not necessarily mean that there is not enough timber in the locality to support the mills. But if the small mill manager does not have access to his own raw material he must face the high cost of acquiring non-owned timber from public or private sources. . The relatively low entrLcost into the lumber business by others. The purchase of a ”belly—up” operation or second— hand equipment can be made with limited capital resources. Mead (30, p. 123) notes that, When demand conditions improve significantly lumber prices increase and milling profits temporarily improve thereby inviting entry into the industry. Small mills enter in response to the invitation, and the favorable operating profit position is quickly erroded away. . The availability of adequate skilled labor. While the number of skilled. workers required is not exceedingly high, they often cannot be replaced. An example is the current shortage of head sawyers and millwrights (maintenance men). Limited sources of capital. Complicated equity or debt financing normally is not available to the small operator. Local bankers and friends are usually the prime sources of funds. . A complicated marketing structure. This includes many 8 different avenues for disposing of the product, each taking a share of the selling price, usually in the form of discounts. 6. The lack of improved managerial techniques. Even when the manager recognizes the need for improvement, additional training and help are often not available to him, or will not fit into his busy schedule. It would be easy to continue the list of difficulties of these businesses in today's world. Even if the manager does cope with these factors, he is forced to ponder the increasing technological gap between his company and his large corporate counterpart. His sporadic and limited excess capital often must go into maintenance, or if he is lucky, into minor production improvements. Bigger firms routinely allocate funds for process sophistication, market expansion and some research and development. And so the gap grows larger and larger between the two. On the surface it is anything but an Optimistic picture, and yet many of these small managers are succeeding. Why? Is it the un- spoken optimism and nature of the leadership? Is it the character of the organization? Do small mills possess advantages of size that offset the power of the larger firm? No group would more appreciate the answers than the managers of the small sawmills themselves. Managerial Time -Use In the search for published studies dealing with how managers use their time, a relatively small number of references could be found. This is not hard to understand. It has only been in recent years that concepts developed by behavioral or management sciences have experienced recognition, testing and success in industry. Carlson's study of Swedish managing directors, Brisley's (6) study of a group of Detroit executives, and Stogdill and Shartle's (39) study of wholesale executives and Navy officers are examples of research that produced data showing percentages of time spent by managers in different activities. In addition some quasi-managerial studies, rang- ing from engineers to office personnel, are available. No time studies concerning sawmill managers in the United States could be found. Moreover, no comparable published research on managerial behavior in any aspect of the Forest Products industry was located. One Swedish study (1) was found concerning supervisory personnel in Forest Units, and one proposal by Battelle Institute (3) to investigate the decision making process of forest managers. Time plays an important role in this research. It was used to measure behavior periods and to search for commonness and differ- ences between managers. An effort will be made to tie empirical behavior data concerning time-use and communication channels to theoretical role requirements of previously established types of 10 leadership. The individual analyzed in this study will be the manager or the owner—manager of the mill, who therefore has the sole responsibility for the direction of his company. No attempt will be made to estab- lish who "should not" or ”will not" be in business in the future, but rather to explore the characteristics of a few of those who are sur- viving with varying degrees of success. The analysis should enable the manager to judge his own activities in light of his own goals. While the temptation is indeed great to draw conclusions about the status ofa_ll managers operating small sawmills, it must be re- membered that the results apply only to the Michigan group. Further regional studies will be necessary to prove the suspicion that the types of managers described in the Michigan sample are also running sawmills in North Carolina, Oregon or Maine. 10m. PART I. METHODOLOGY Random Interval III. Work Sampling II. Technique II. THE SAMPLING PLANS Two distinctly different sampling techniques were used in this research. The first, used to select the managers for the study, was a one-time-only process. The second, used to sample the activities of the mangers, was a continuous process, carried on for a specified pe riod of time. Selection of Managers and Their Mills A list of all the sawmills located within a 100 mile radius of Lansing, Michigan, was compiled using the newly published Michigan Directory of Primary Wood Usirg Plants (31). The size of the mill, precise location, county, and ownership were verified with assistance from the Michigan Department of Conservation and the Cooperative Extension Service. Mills beyond the 100 mile radius were not considered for the study because of the practical constraints of driving time, and be- cause there were enough mills within that area to provide an adequate sample. Although the study was limited to mills producing five mil- lion board feet or less per year, there were 140 mills within the radius from which to select the sample. After further investigation, 11 12 mills producing less than one million board feet per year were elimi- nated from the study. They do not operate on a regular basis, they usually are low manpower operations, and often these mills are not permanent. I In fact, some could not even be found! The mills remaining for consideration then fell into Michigan classifications C and D (see Legend, Figure 1). Over 30 telephone calls were made to mills scattered throughout the 100 mile radius, and each manager who showed any interest in participation was visited. In the final analysis, the list was randomly thinned by the managers themselves; those who were willing to make a commitment to a considerable amount of work, with the hope of learning more about themselves as managers. The vital statistics of the 11 companies to be studied were as follows: Sowner - managers 5 co-owner - managers 1 manager (absentee owner) 11 total participating managers (or key decision makers) 4 Class C mills out of a state total of 14 7 Class D mills out of a state total of 74 11 Total mills out of a state total of 88 Class C and D mills i ’7 The sample was 12.5% of the managers operating plants within the state in the one to five million board feet per year production A permanent mill usually exhibits a concrete foundation, a planned site near road and /or rail, and a pattern of regular operation. In contrast, many small mills are portable, skid or tire mounted, and run according to the immediate needs of the operator and his local market. //// / LEGEND 0 Mill location - 11 managers in study * Lansing, Michigan \\ // SAWMILL CIASSII-‘ICATION (Production Classes, Annual Basis) A Over 7, 500 MM. 8 5,001 to 7, 500 be. c 3,001 to 5,000 be. D 1,001 to 3,000 m. r. 501 to 1,000 be. II“ \\ \\\\ \\\\ \.{\\ F 100 to 500 MM. F- Under 100 MM. Figure 1. Location of 11 sawmills in Michigan study. 14 range.1 The map in Figure l locates the 11 sawmills. Standard in— formation about each plant can be found in Appendix A. . Z . Work Sampling as a Technique Even before the plants had been selected, considerable time had been spent determining the manage rial activities that were to be measured, and selecting the sampling methods to be used. The deci- sion was ultimately made to use a modified form of work sampling. The continuous time study, another method of work measure- ment, has most commonly been used with production line workers. Similarly, the "11 and 4” system3 employed by supervisory people at Oldsmobile, and the executive-time-log advocated by Drucker (16), are two modified ways to account for every minute of time. While less precise than the stop-watch analysis, these are still forms of continuous time study. What is work sampligg and how can it be used? The work sampling technique has been employed for some time as a method of measuring work. Barnes (2, p. 11) states, ”Work 1Since 1968 one of the Class C mills has increased production to 7MM/year, moving into the lower range of the Medium sized mills. 2 The development of work sampling is attributed to L. H. C. Tripplet in his 1927 study of lost time in weaving. 3Al: 11 o'clock and at 4 o'clock, the participant records from memory his activities for the day. 15 sampling is based upon the laws of probability. " The statistician calls it random sampling: the larger the sample, the greater the degree of accuracy in the predictions. The application of this method seems to be in three distinctly different areas. One use concentrates on forecasting, maintaining, and improving the quality of products, or quality control. Another use establishes or sets work standards, or time standards, for a job. One of the earliest and simplest examples of this was called the ”ratio-delay" method; it measured the ratio of work-time to delay-time of machines or workers. Conway (12) pro- vides a recent forestry application, pinpointing delays in logging operations. A third use is an increasing tendency to analyze different groups of people, such as school teachers (2), engineers (40), and missile workers (36). When used in this way, concentration is more on what people do and less on machines and production schedules. Gradually the idea of research on managers is being accepted, although the total number of managerial time studies is small, even when methods other than work sampling are included. Advantages and limitations of work sampling Traditional work sampling, where an outside observer makes random observations of his subject at work, possesses several key advantages over the continuous time study (2). Some advantages 16 which apply to this study are: 1. Work sampling may be preferred by those who are being studied; some people do not like being continuously observed. . Work sampling is less likely to cause the observed to change his routine or alter his behavior. Little observer training is required and the fatigue factor of the observer is much less. . Observations can be spread over days, weeks, or months. thereby giving a fairer view of activity. . The study can be interrupted, stopped and restarted without seriously affecting the results. . The degree of accuracy can be chosen in advance by prede- termining the sample size. . It usually requires less time to calculate the results of work sampling, especially with the use of computers. Some limitations of traditional work sampling, in terms of this study are: 1. In situations requiring one observer for one operator, it is uneconomical. . The operator may change his work pattern when being watched; the results would then have little value. . Sample size has to be quite large if a high precision estimate is required. . Work sampling is difficult to explain to the participants. 17 Since this study is concerned with managerial behavior, two more limitations of the traditional work sampling method must be considered. A manager's job is both versatile and mobile. An ob- server has a difficult time evaluating activities such as talking, read- ing or thinking (7) and he has the problem of following the manager to a variety of locations. An alternative: self observation system Faced with the limitations of the traditional work sampling approach, an alternative method was suggested by Dr. Miles Martin, Department of Communications, Michigan State University. A self observation system incorporating a novel Random Alarm Mechanism had been tried with success by Martin (26). Its major advantage: the observer and the observed were the same person. Recently, several other work sampling studies have been con- ducted using some form of the self observational technique. White (40), in a pilot study of engineers, used a low power, centrally located radio transmitter which emitted. a tone and then a voice message to each participant carrying a vest pocket receiver. The tones arrived randomly. Carroll and Taylor (7) tested the validity of the self observation- al central signalling method. against traditional work sampling with clerks in an office. Overhead. lights were flicked on and off at random 1Currently at Syracuse University, Syracuse, New York. 18 times as a signal for the workers to check an activity sheet. An out- side observer made another set of observations. The differences be- tween the methods ranged from .1 to 3. 3%, depending upon the activity. While this variation would not be accurate enough for setting job standards, they felt that a self observational technique would be the most realistic approach to studying time allocations ofpersonnel in managerial positions. Even with its own set of limitations, for this study the self observation system could overcome most of the disadvantages of traditional work sampling, and still retain the advantages that work sampling has over the continuous time study. 1. The cost of conducting the study would be substantially re- duced. Either of the other methods would have required 276 man days devoted to gathering data alone by an outside observer. 2. Although the possibility of behavior change cannot be ruled out, the participant would not be influenced to change his work pattern in the same way he might if he were being watched. Some self-reporting bias could occur, however, unless he was guaranteed anonymity of his recording. 3. The sample size could be adequately large because outside observers would not be required to gather the data. 4. Activities such as "think time' and "getting information about a problem" could be self-evaluated if meaningful categories 19 of activity were provided for the participant. 5. Freedom of movement would be possible. Because of the location of sawmills, a random signal transmitted by radio or by telephone would have been prohibitive in terms of dis— tance and cost. The compact individualized Random Alarm Mechanism could overcome this obstacle. Pilot 8 tud ie 3 Twelve Random Alarm Mechanisms were obtained on loan from Merck, Sharp and Dohme on a trial basis. At the time of their acqui- sition, they were the only devices being manufactured in the United States with a set average signal ratel considered acceptable to manag- ers.. Several people in the Forest Products Department at Michigan State University carried a unit for two weeks, recording alarm times and activities. The units performed according to their specifications, but several shortcomings were evident in light of the intended use. These will be discussed in Chapter III. Perhaps one of the most important aspects of this study was the design and building of the new random alarm instrument to meet the specific needs of sawmill managers and their surroundings. Twelve new devices were built by Thomas H. Charters, an electrical engineer, according to desired specifications. A bench test conducted on 1One of the tested Random Alarm Mechanisms had an average rate of one signal every 151 minutes. 20 instrument Serial Number 81 constituted the second pilot study. Chapter III is devoted entirely to the new Random Interval Timer. Before enough of the new instruments were available to begin the general study, one manager agreed to use the first device on a trial basis. It was set to emit an average of one signal every 60 min- utes. This manager's response to the trial period indicated that such a large number of signals per day would irritate many managers and the rate should be reduced. As the rest of the Random Interval Timers arrived and were put into the field, they were set to signal an average of once every 80 minutes (or :l:8 signals per day), which seemed to be an acceptable rate for this group of men. General Work Sampling Procedures and Coverage If interest was shown by a mill manager during the initial tele- phone contact, a standard procedure of explanation developed by the writer was followed. This involved traveling to the mill, explaining the project and, if the manager. was still interested, leaving an in- struction sheet, Random Interval Timer and data cards. The com- plete instruction sheet is shown in Appendix B. A minimum of four trips was made to each participant to gather extensive interview information, replenish data cards, and service instruments if needed. It was stressed throughout the study that individual time-use results would remain anonymous, and therefore numbers were used to designate each participant and his firm. 21 Data Card de sign Key activities as they pertain to managers were listed on the Data Card, utilizing the few previous studies available. The first devised form was too detailed and the second was confusing. The sample card shown in Figure 2 represents the final product. The work classifications had to be meaningful to a sawmill manager and easy to record. Four of the listed activities attempted to uncover the amount of time spent in the decision making process. This decision making group included: Receiving a problem: any kind of problem, from broken equip- ment to a dissatisfied customer. Getting information consultation with others, reading and about a problem: writing for available information. Deciding action consultation with others, "think time" about a problem: and "creative time"; includes weighing alternatives. Giving instructions communicating how and what shall be about a problem: done after the decision has been reached. The rest of the activities were self-explanatory. Each card was designed to be used for one day (up to 15 signals), and when folded would fit into a shirt pocket. Space was provided to record START time (turning on the instrument) and OFF time; this information was 22 (Back side) Mill No. 5’8 Time Month [1% Spent Day 25 M®w TH E s RIT Prior Card No. 2 Time RIT Start 7:45 Training . 2’ Receiving a poblem 8' ’0 5 i Getting info about 2 8:37 2 problem I l 3 Deciding action 3 9’03 5 about problem 4 /0; /0 . (think time too) 5 /2:'// 3 (3ng imtructiom / about a problem 6 AWfi 25 I? Routine wprk 7 am 0 Job mtemews __Correspondence 8 31.45 7 3? Travel . 2 Break time (personal 9 Z/5 Z also) 10 Q Other (describe) 11 12 13 14 I EVENT 3 - I .121... ...- l I] I]. ///:m {A .I.‘ .4 m ”Mi? I EVENT 4- ./ ’1..../ - I I I I 1".4-A4m/41 /. A”. 4 I . . 4’1“, ’ 1.. h“ ’71.. 4 41]-! M (Front side) Figure 2. EVENT 11- 15va 12- EVENT 13- EVENT 14- EVENT 1 5- Example of Data Card. 23 important in establishing interval times between signals. Each time the Random Interval Timer emitted a tone, the man.— ager would record the time under RIT Time and then write down, under EVENT, what he was doing and with whom he was communi- cating, if applicable. A special column was provided for recording the length of time he had been engaged in that particular activity prior to the signal. Finally, the user was to classify by event number the activity designation which best described the event; TRAINING, ROUTINE WORK, e tc. Proposed vs. actual pattern of coverage Four sampling plans had been designed in the early planning stages, each with a different number of total observations. In all experiments the researcher is caught between the perfect design and the minimum acceptable to provide useful data. This study was no different; the final group of participants, the available time, arrival ofnew Random Interval Timers and travel cost contributed to a feasible compromise plan. Table 2 shows the best proposed plan and the actual coverage as it occurred. Numbers 1 through 12 corres- pond to the mill manager's code designation. Each manager's observations included different weeks in differ- ent months to avoid any bias due to a concentration of unusual circum- stances. No data was collected at any mills during the first week in July; traditionally sawmills are closed that week for repairs and the :8: :.oH.o :4: 24 Z < Z . . . . . . . . . chums/u. <\ \ o w m N. m m m m w 010$ 2.010 :6“ :.oH.o mnoflmog HmuOu mHN H + HON Uh QHWE HH . . a a a a q a . a . NAHUM: N oNH w.~..o.N.H @wa nova :32 um5\uom£o+SN . . >mm32 mnmwmcflz 3:05800 faoz m w m N a £003 onSoc/nomno mHNN n Z mnowmnmz : mwnuocwoo assholes H33 oomN n nomno OONx mama NH m.N.~ N~.:.oH o.w.> 06$ umsmoxw nmctnmmno OON u nw8\>mp\uomo.o . . . . . . . . Sxowumsse wastes omumxmmav e m w m NH 2 Z S o w s 33. 43:08 some ion? 0 .wN o.m.v MIN; N128: 92:. 280:6 m E 3183 newscast comm 2 .2 .2 m .w .s o .m .w m .N .z .32 nmwdqg muGoEEOO £332 1.. m N a x06? mnowumc’uomgo oovN n Z numwdcmz NH swam homomoxm .3303 was mauGoE >0. mnommnme mo owmnocyoo Hagen paw Goa pomomobm .N 3an. 25 July Fourth holiday. A composite comparison of monthly, weekly and daily observa- tions is shown in Table 3. The observations become more evenly distributed as the units become smaller, progressing from months to weeks to days. The first five days show a relatively even coverage. If, for instance, the entire group of managers spent Thursday after- noons buying timber, there were approximately as many observations (427) occurring on Thurdays to capture this behavior as there were on any other days of the week. Table 3. Total distribution of observations (N) by months, weeks and days. Weeks of Days of Months N All Months N All Weeks N May 189 lst 392 Mon 404 June 465 2nd 561 Tue 429 July 913 3rd 531 Wed 435 Aug 648 4th 496 Thur 427 5th 235 Fri 420 Sat 1001 Total N 2215 2215 2215 1Observations for Saturday are low because some man- agers operated only half a day. The coverage by hours of the day was also important. The dis- tribution of observations should be nearly equal for each hour of the day if the managers have not turned the Random Interval Timer off during the day or missed any signals. 26 In Figure 3 the total distribution of signals is shown for each hour of the day. Different START and OFF times of the managers explains the lower number of observations for the early morning and late evening hours. They had been instructed to switch on when they left their homes and to leave the instrument on until 9:30 p.m. if any company business was being conducted. However, after 4 p.m. a few of the early starters turned off their devices, and the rest of the managers continued to turn off at various times up to 10 p.m. The hours between the dotted lines represent the average 10 hour day for most of these managers. The upper bold line (and upper figure) represents the total of all observations recorded during each hour. The lower checkered line (and lower figure) represents the remaining total after Saturdays have been subtracted, because some managers did not work a full day. It is not possible to explain with any certainty the higher than normal deviation in the 12 and 1 o'clock hours. Some operators may have turned the devices off after being annoyed during lunch and for- gotten to turn them on again until later. Examples of the distribution of signals (or observations) for three of the instruments is shown in Figure 4. Here, as in Figure 3, the lower line and numbers indicate the subtraction of Saturdays. While it is felt that any error occurring was most likely human, the instruments would have to be run under varying conditions of move- ment and temperature to prove that some error was not caused by the 27 .Amnowmcma :v >m© of mo pnos fiomo wcwndp Umpnoomn mfimcmfim ong .m ohsmfim >40 MI» “.0 mmDOI .2 .n. . 2 .4 m m N. m n v m N _ N. _. O. m m h m 0. mm s E N0 m0. mm. NV. ON. N: sh. V0. mm. *0. NO. mm. mm. .1 :1 0! mt 0N... cm. 00. .vON mm. mm. m5. 0w. 1 ....._ _.....q . .l u n... - . _ . _ _ _ 1 _ _ . . .A >40 )v. 0 On 00. on. CON S'IVNQIS :10 BBBWON NUMBER OF SIGNALS 28 40- NOJ - INSTRUMENT NO. 84 I— T" ,_ F Fl annuazuasnazs 2I 22] 24 37 38 2I 26 3 as) 25 0 ‘OF no.6 INSTRUMENT N0.8I8 20' I. fl 7' T- r V "F“ . 22°21- 2' 3231.12 1.2 '3 0 I5 19 20 2| l9 Is 2| I8 I5 I3 40)- 'No.Io INSTRUMENT N0. 817 20r— ! F g F- F m gazugnasazsnu o 16 l7 IS 23 I5 24 l9 24 I4 II I I I I 4 AJ‘. 8 9 O I 2 2 3 5 RM. HOURS OF DAY Figure 4. Examples of the distribution of signals per hour of three managers. 29 instrument. A larger individual manager sample would improve the distribution if the possibility of human and instrument error was eliminated. However, the prime concern in this study is the behavior of the entire group, and of the sub-groups, and therefore utilizes pooled data. Accuracy of Work Sampling Measurement One of the difficulties any researcher faces is that of determin- ing a satisfactory number of observations. For sawmill managers, there were no previous studies to indicate the percentage of occur- rence each activity would occupy out of the total observations. The decision was made in the beginning to use a confidence level of 95%; in worksampling this is the most widely used. Barnes (2, p. 13), explaining the meaning in simple terms says, "This means that one is confident that 95% of the time the random observations will represent the facts, and 5% of the time they will not. " Based upon the 95% con- fidence level, the expression for determining the accuracy for a given number of observations out of the total is: s(p) = $1.96 P'—'1\I‘-E' where S-the degree of accuracy 1Degree of accuracy is the work sampling term comparable to the statistical Standard Error of the Mean. 30 p- the percentage occurrence of the activity being measured; p is the percentage of the total number, N. N- the total number of occurrences (signals) in the sample. 1.96- establishes the left and right critical regions of the normal curve, under which 95% of the area is represented (or the 95% confidence level). The relationship between the desired degree of accuracy and percentage occurrence of an activity can be seen in Table 4. The higher the percentage an activity occupies of the total sample, the greater the degree of accuracy. As an example, ROUTINE WORK occupied 30. 33% of the total observations. Using the above formula, the accuracy is $6. 81% o_f 30. 33%. Down the list, JOB INTERVIEWS made up 0. 98 (or 1%) of the total observations and the degree of accuracy slips to 44.40% o_f 1%. By multiplying the percentage of occurrence by the degree of accuracy the Absolute Error and True Value range can be found. The accuracy of 4:6. 81% of 30.33% equals the Absolute Error of :tZ. 04%. This is another way of saying that for ROUTINE WORK the True Value lies between 28. 29% and 32. 37% (30.33% :i: 2.04%) and. there is the probability that 95% of the time (confidence level) this is a true repre- sentation. The practical question arises: Is it better to have a high degree of accuracy for activities that represent from 1% to 10% of the total 31 - - - s8 .2: SE monH<>mMmmo HEHHCSVQ AMHHZH mob. owN -NN; memo oon 54 mm EHJmOMQHmOmm oo .w .00 .N mm .0 mm .NN ww . m 3. HOZHQZOnHmMMMOO owe Iwmgv Ho; wodfi om.m mo“ A mNA mmél mm.m m3 ZOHHO< OZHQHOMQ MN.:-sm.w mm; 34.2 cod :1 AW>§H wo.:-oo.w cm; 2.2 Nm.oH mm: ”52:. M nonum 330mn< 4393004 mo omummfl H.308 mo .x. .mnO mo 33304 cash. as $3 1.5m .oZ .mnowmcmg HZEBNm : mo mofigflom vantage 05 MOM 03.3 43.945064 .4. 3an 32 (for which a tremendously large sample size would be necessary), or is it better to keep the Absolute Error below a reasonable mini- mum for the entire study? The answer lies in the type of research. For a managerial study, a variation of i2. 5% Absolute Error is acceptable. To lower the Absolute Error to 11% for ROUTINE WORK, thereby narrowing the True Value range slightly, would have required approximately 8200 observations. The additional cost of expanding the study to four times its size is not warranted for the sake of saying that 30.33% could have been 29.33% or 31.33% instead of 28. 29% or 32. 37%! III. 'IHE RANDOM INTERVAL TIMER History of the Instrument In a developmental sense, the history of the Random Interval Timer or Random Alarm Mechanism is most fascinating. The com- plexities of developing a small electronic device that will produce an audible signal at random moments in time cannot be underestimated. The first prototype was developed by Stuart Cooke of Case Institute of Technology in 1960 for use by their Operations Research Group in conducting a nationwide study involving chemists and physi- cists (26). The basic design was simple and ingenious, but required some skill on the part of the carrier in the ”re-set" operation by counting ten flashes on two small neon lights and returning the switch to the "run" position. Its major drawbacks were: (1) it would not signal until a lag period of two or three hours had elapsed after it was initially started, and (2) its output was dependent on the position of two mercury tilt-switches induced by motion. A commercially built unit called RAM-l1 came on the market in the mid-nineteen sixties and offers some improvements over the 1RAM-l, Random Sampler, Electronic Ideas, Inc., Wyncote, Pennsylvania. 33 34 original design. It is compact and has only one small fiber wheel for the operator to manipulate. Since the instrument is built to produce one specified average number of signals, its lack of versatility limits its possible applications. There is also a more serious problem; the device could cease to be a truely random timer if the operator forgets to turn the fiber wheel, which is the heart of its unpredictability. With these shortcomings in mind, Mr. Thomas H. Charters, a consulting electrical engineer, was asked to design a new instrument. General Requirements for Design of the Instrument It was suggested that for this study involving sawmill managers the desirable features of the new device should include: 1. Self-contained circuitry that did not require the operator to turn wheels or switches after each audible alarm. 2. A choice of several average signal rates so that the annoy- ance factor could be accommodated to the user. 3. A built-in speaker and external earphone jack with a volume control. This would allow the user to adapt the device to various noise levels found around heavy machinery without being embarrassing in quiet environments. 4. A reasonable size and weight. 5. Belt straps for concealed and convenient carrying. 35 Technical Requirements for Design of the Instrument The technical objective was to produce a sampling device which would sample uniformly in time but in such a manner that the user could not anticipate the signal. The uniformity in time specification arose from the logic that each hour of the day should be sampled with an approximately equal number of observations. Satisfaction of this requirement could have been achieved by selecting a reading every 30 minutes and asking the manager to record what he was doing at 8:00 a. m. , 8:30 a.m. , 9:00 a.m. and so on. Three difficulties arise with this technique. (1) It requires the manager to remember to look at his watch and write down his observations. (2) If he forgets an observation and. his memory is unreliable, the results could be biased as he tries to re- call his actions. (3) His normal behavior pattern could be changed if he began to anticipate the established. times. Another way to achieve uniformity could have been the use of a random time table. By taking a large number of observations, each hour of the day would be fairly represented. This technique would. produce all the difficulties mentioned in the previous paragraph since the random times would have to be written down, carried and re- ferred to. The Random Interval Timer solves these three difficulties and also satisfies the second portion of the technical objective. The user 36 cannot anticipate the signal. Uniformity over time should be achieved if the number of observations is large enough and the user does not turn the Random Interval Timer off and on during the period of study. The theoretical model The requirements mentioned in the preceding paragraphs are best fulfilled by the Poisson Model or Poisson process. The approach toward choosing a mathematical model for this study is reversed from most real-world situations. Normally the researcher is confronted with a collection of data that he believes be- haves according to a particular model such as arrivals of autos at toll booths or coal mining disasters. If, by analysis, the historical evidence compares well with that model, he can make good approxi- mations about the future from that model, such as the maximum back up of vehicles or time between mining accidents (35). In the case of the Random Interval Timer, a theoretical model was chosen whose behavior best fulfilled theiechnical requirements without neglecting the realities of design. It can be used. to collect data for a wide range of studies if the conditions are thought to be the same as the model. The Random Interval Timer generates events (signals) according to a model rather than the researcher recording events as they occur and then finding the best-fit model. Finally, the instrument can be tested against its own theoretical model to see if it is operating according to design specifications. 37 The main features incorporated in the definition (35) of the Poisson process are: l. l The process is stable. The mean rate of occurrence, p. must be constant if the series is to be random. The probability of a signal occurring in the inte rval of time t to t+At isalways pAt. . Any Chance of a signal (event) occurring in the interval of time t to t + At is independent of all prior signals (14). Two or more signals cannot occur simultaneously. There will always be some measurable interval between any pair of signals. The signals are seemingly'uncontrolled and unpredictable. but over a long period of time the total number of events are predictable (equal to pT, where p. = the mean rate of occurrence and T = total elapsed time). The above five features make up the main definition of the pro- C888. Three additional characteristics are included because of their graphical application. 6. The signals have a Poisson distribution. The probability (chance of occurrence) for a given number of signals in a given period of time may be calculated. from the expression: 1Also called the ”average arrival rate. " "rate of occurrence. " ”a reciprocal of time" depending on the source of reference. it is not to be confused with M. the mean of the Poisson distribution; M = [-Lt- -pt k P(k) = —’-‘——e k[( t) where: P(k) = the probability of obtaining exactly k signals (events) it = mean rate of occurrence k = number of signals (events) time ('7- ll 7. Distribution of time intervals. At. between signals will have an exponential distribution with the mean interval equal to l/p and defined by the expressions: P(t) = we“) where: P(t) is the probability of obtaining an interval length of value t. a point on the smooth expo- nential curve. A more useful expression is to construct a grouped frequency distribution. or histogram. with At wide bars: AtP(t) = (Mae-M) where: AtP(t) is the probability that an interval of time, At, is centered upon mid-point value t. It is this characteristic exponential distribution of the time intervals that reassured the researcher that the users of the 39 instrument would be prevented from anticipating the next signal. 8. Using a large enough sample size, the distribution of signals over time (hours of the day in this study) are uniform. With the theoretical Poisson model in mind, it was possible to design a new type of random alarm mechanism. The new Random Interval Timer Mr. Thomas H. Charters, operating within the general and technical constraints. invented and applied for a patent (9) on the instrument shown in Figure 5 and Figure 6. In Figure 5, the numbers apply to circuitry and the letters to wave—forms. There are two pulse inputs into a coincidence detector, 14. When these pulses coincide. an audible output is produced. The source of one of the pulse inputs, and a critical feature of this device, is a random noise generator, 10, producing a noise sig— nal A that modulates the output of the rectangular wave generator. 12. The rectangular wave generator is capable of producing a train of rectangular waves, B, itself, if it is not affected by the A input. The noise input is effective for either advancing or retarding the occurrence of each pulse output. As a re sult, an output signal. B. is produced comprising negative-going pulses occurring at sub- stantially random times into the coincidence detector, 14. In design- ing the circuit, the duration of each negative-going pulse, B, was 40 H—ems.——H 3-|zmin.+-a-1 no :2 B) I -' l4 E/L f 2 K. : RANDOM RECTANGULAR COINCIDENCE SAMPLING NOISE WAVE DETECTOR RAMP GENERATOR GENERATOR GATE 7,-H A OSCILLATOR , hon UUU Figure 5. Block diagram of new Random Interval Timer1 and wave- forms. lFrom U. S. Patent Office Application Serial No. 805-339. 41 Figure 6. Photograph of the Random Interval Timer. 42 established at approximately ten percent of the period between pulse occurrences (9). Therefore, there is about a ten percent chance that the other given wave-form. output G. will coincide with wave— form B in the coincidence detector, 14. At this point the device can best be described as a systematic Bernoulli process. The sampling principle is analogous to flipping a coin with one head and nine tails. On the average. one out of ten tosses will produce a head, or signal in this case. If a large number of these repeated Bernoulli trials were plotted with the established probability of coincidence of approximately 0. 1 at each sampling. the binomial distribution shown in Figure 7 would result from calculations in Table 5. Where the Bernoulli process is characterized by the prob- ability of success (signal) on any trial, the Poisson process is characterized by the expected number of successes (sig- nals) per unit of space (time) (35, p. 212). The second pulse input with two time delay circuits provides a signal period on the order of minutes. The second input is provided from a pulse generator comprising a timing ramp, 16. a sampling ramp gate. 18, and a sampling ramp, 20. The timing ramp allows the user to select a sample period in advance. There are eight choices. or strap settings, ranging from 3 to 12 minutes. Since wave-form G will coincide with wave-form B on the average of once every 10 samples. it becomes a matter of multiplying by ten to find. the average signal period. 43 Example: Strap setting No. 3 has an 8.0 minute sample period. A probability of coincidence of approximately 0. l at each sampling will yield an average of one alarm every 80 minutes (10 x 8.0 minutes). The sampling ramp gate, 18, and the sampling ramp, 20. merely take the relatively long duration wave-form C (several minutes) and generates a pulse for input into the coincidence detector. In other words. the second time delay circuit, 20, takes the output of 18 and produces a pulse, G, in a very short period of time. This prevents pulse G from being influenced by wave-form B. The dotted line, 22, in Figure 5 shows all the wave-forms generated in the instrument and the instance of a coincidence between pulses B and, G. Pulse H then triggers the oscillator gate, 24, which allows the oscillator, 26, to operate. generating an output signal tone .1 in the speaker, 28. The relationship between the binomial probability law to reduce approximately to the Poisson probability law is seen in Figure 7 and Table 5. These curves will remain the same regardless of the time setting because even though the average signal rates may change. the probability of coincidence remains 0. 1. Therefore, the probability of obtaining exactly one signal is greatest when the elapsed time is equal to ten times the sampling period of the timing ramp, 16. There are three innovations in this instrument: 1. The detection of coincidence between two separate pulses. 44 0.4 .. /"‘\ m Binomial Distribution .- \ O | K n -K - /° p= "‘pU'p) K.‘ ( n — K) ! 03- l/ .._, ‘ Poisson .. $3 Distribution o.2L( 0.: " Data from RIT No. 52v 5. ———L_,O Probability of obtaining exactly one signal, p/K) o 1 l 1 l l l l I 0 I 2 3 4 pt or np Figure '7. Poisson. binomial probability distributions of the Random Interval Timer. 45 Table 5. The mathematical comparison of the binomial and Poisson probability distributiom. Binomial Poisson n = number of samples 11 = mean rate of occurrence or average signal . rate k = number of successes, Signals or events k = b f it . P = Probability of success at each sample num er 0 successes, 51811315 01‘ events (l-P) = probability of failure at each sample P = pmbabfllw Of success t =time If k = 1, then P(k) = the probability that one signal, and one only, will occur in time t. The following table was constructed and the results plotted in Figure 7, Bod: eacpnessions are simplified when k = 1. Binomial pt or 11 P22?! Paon 1 h g .2) ' n? w en n-1 n-1 4“ "Pt n0 9: = n? P = 0.1 n-l (o. 9) n(0. 9) e e (1.1.1:) 10 0. 20 2 1 0. SEC 1. 80 0. 819 0.164 0. 180 0. 50 S 4 0. 656 3 28 0. 607 ‘ 0. 303 0. 328 0. 70 7 6 0. 531 3. 71 o. 497 0.348 0. 371 1.00 10 9 0. 387 3. 87 0. 368 0. 368 0. 387 1. 20 12 11 0. 314 3. 77 o. 301 0. 361 o. 377 1. 50 15 14 0. 229 3. 44 0. 223 0. 334 0. 344 2. 00 20 19 0. 135 2. 70 0. 135 0. 270 0. 270 3.00 30 29 0. 047 1. 41 0.050 0. 150 0. 141 4. 00 40 39 0. 016 0. 66 0. 018 0.073 0. 066 Note: This curve is universal--it will apply for all signal rates selected in the Random Interval Timer. The peak occurs at pt = nP = 1 in = 1 when nP = 1 where P = 0. 1 '. 1 1 n = R =0-1- = 10 which means that the probability of getting one signal is greatest when circuit takes 10 samples. 46 one random and one periodic. The use of the two ramps to insure the independence of pulses in the coincidence detector. The controlled probability of coincidence, in this case small enough so that the output signals substantially satisfy the definitions of the Poisson process. The present circuit is automatic and continues to produce output signals having a Poisson distribution without resetting or other human intervention, thereby rendering the output distribution substantially free of error. Te sting the Instrument The output of this device can easily be tested against the Poisson distribution. All of the tests referred to in this section were made on the first production model, Serial Number 81. The instrument was programmed for an average of one signal every 25 minutes. Three characteristic curves of the Poisson process are shown in Figure 8. 1. Figure 8a: The Probability of Obtaining Exactly One Occur- rence (signal). Other curves for k = 2, k = 3, k = n have been omitted. Figure 8b: Percentage of Occurrences. This distribution shows the percentage of waiting times or intervals between signals. and is exponential for a Poisson process. This 47 K5 3 S *1 N k u ‘3 E o. k ‘6 0: § 28 ‘0‘ -’uf : NT. =,ut (e ) 2‘ g‘ When K :I Q 3.\ 33: g u 0. V 11‘. .111. Wayne; b. 9.00 = 100 A A1 (3 ) O K ° 0 us... to " N Q: E 50m”; V E k 10.. R § 0: ‘04 § E 8 s.) Q ‘i‘ c. o A A A A ‘ . l I l l I l I 1 'r - , - - - .- I 4 0.2 .4 3.8 I Zflt 3 [Ailatlaflatlafl 0 51052025 . 50 . 75 '00 "mam, minutes C. 100 50 CUMULA T I VE DISTRIBUTION FUNCTION, (CD. I". I, % Figure 8. Three characteristic curves of the Poisson process. 48 histogram is helpful in detecting deviations of the observed data. The theoretical curve is also shown. 3. Figure 8c: Cumulative Distribution Function. This theoreti- cal curve was the most useful of the three in this study as the empirical data could be compared to it using a computer program of a one-sample Kolmogorov-Smirnov test of fit (32). With the three curves of Figure 8 as a base of comparison, a closer look was taken at test instrument Serial Number 81. The re- sults already plotted from Table 5 in Figure 7 produce an accurate representation of Figure 8a. Still another method of plotting is shown in Figure 9 using Table 6 for the data. By actually counting the number of signals that occurred in every 24 minute period along the continuum of total time the instrument ran, the percentage distribution of signals per set time period can be plotted. An example in Table 6 shows that one signal occurred in 34 out of 89 of the 24 minute time periods. In Figure 9, the observed data is plotted against the theoretical values. A good correlation exists for such a short bench test. The fit shows that the actual signal occurrences have a Poisson distribution. If the operating device is performing as designed. then given the Poisson process. the distribution of time intervals between signals is exponential. Figure 8b shows the theoretical distribution of the time intervals with the mean rate of occurrence set in the instrument 49 P = . l l signal/24 minutes mean rate of occurrence 1: n t = 24 minute time period T = 36 hours 4.0 —_ N = 89 signals 0"" 3 7 . 3OJ / \\ --.-— Theoretical o g \\ % Actual 3 \ 5 \ 3 20 - '-H I/ I t? a . / a: .C. / U 10 “ / \ 5° E / \ / \ / " / ' . 0, _/ in \1. 'r'. 1.: 0 1 Z 3 4 5 6 No. of signals/Z4 minute period Figure 9. Distribution of signals per time period--actual vs. theoretical for SN 81. 50 «363 mm mm 92: coo; o . o c To Moo. 0 o .o o o m .o moo . m fl 4 v A m 4 mac . v N. .m m~ m To Hoe. m N.oH Tm : v.2 «.3. N 06m. mm mm mom mom. H 0 Km 0 mm w on mom. 0 00:92:000 mHmomfim mo .02 mqwhu0000 wfiun5000 mo ._ x No 0052.6 Hmfloouoozfi 209.qu mo .02 oocmsu .x. a HmoflouoonH 1330.80an x o n o .2: mm mm 0 0| ol 0 H A m d m H A v a v m 1v NH v m N .oH mm 2 N m .wm mm mm H w .mm o Nm 0 002935000 mfiocwwm $393000 .063qu o\ .02 Av: uo 00:02.6 mo .02 3304 mo .02 13304 oozed 58 ¢N\ .oZ mfimcmfim .x. Hood—0.4 Aw 2m HON amufiouoosu .m> #9300 oofluom 053 you mfimawdm mo Gofioohuomwfl .e 3an 51 1 PL = -2—5- . The Percentage of Occurrences column of Table 7 was at used to plot the theoretical curve with a grouping time interval of five minutes. If the actual number of signal periods (interval in minutes between signals) is plotted against the theoretical number expected. any serious lack of fit should be detected. Figure 10 represents a histogram of SN 81 in exponential form. Notice the slight oversam— pling in the short times and under sampling in the log times (except for 30 and 60 minutes). The best explanation possible is that the total time the instrument was run was too short for such an extensive statistical comparison. In the actual study these devices were run for weeks so that the total number of signals was substantial. Still another way to look at the distribution of interval times. given the Poisson process. is to plot the curve of the exponential Ft. Then for a particular p (-l- or 0. 04), one function P(t) = pe 25 can determine t ‘. t t . .t so that the area under the curve 1 2’3". 10’ is divided into equal parts. Using it = 0.04 we solve for t from the basic expression: t_ t— -t -t -t Spepdt=-ep l=-ep+l O 0 If the number of occurrences within these time intervals is recorded. a visual evaluation can be made on the goodness-of-fit. The 52 Table 7. The calculated percentage of occurrences for time intervals (between signals) in the Poisson process-actual vs. theoretical for test instrument SN 81. Total elapsed time (T) SN 81 was run = 36 hours Total signals N = 89 Mean rate of occurrence p. =1/24 . . .. set it = 1/25 Mean imerval l/p. = 24 minutes . . . . . . set Up. = 25 minutes Percentage Time Interval p. 1'. Of Theoretical Cum. A t Occurrences , of No. of Actual Disc. in 5 Minute (P. O. O) . , (1) A t _ p. t _P‘ t Occurrences No. of Function Brackets Midpoint: e pA t 100(pAt)e N x P. O. 0. Occurences 96 0-5 0. 100 . 905 . 2 18. 10% 16. 1 20 9. 5 5-10 0. 300 . 741 . 2 14. 82 13. 2 14 25. 9 10-15 0.500 .607 .2 12. 14 10.8 12 39.3 15-20 0. 700 . 497 . 2 9. 94 8. 4 9 50. 3 20-25 0. 900 . 407 . 2 8. 14 7. 2 6 59. 3 25-30 1. 100 . 333 . 2 6. 66 5. 9 1 66. 6 30-35 1. 300 . 273 . 2 S. 46 4. 9 6 72. 7 35-40 1.500 .223 .2 4.46 3.9 3 77.7 0-45 1.700 .183 .2 3.66 3.2 2 81.7 45-50 1.900 .150 .2 3.00 2.7 3 85.0 50-55 2. 100 . 122 . 2 2 44 2. 1 2 87. 8 55-60 2 300 .100 .2 2.00 1.8 1 ”.0 60-65 2.500 .082 .2 1.64 1.5 4 91.8 65-70 2. 700 . 067 . 2 1. 34 1. 2 1 93. 3 70-75 2. 900 . 055 . 2 1. 10 1.0 1 94. 5 75-80 3. 100 . 045 . 2 . 90 . 8 1 95. 5 80-85 3. 300 . 037 . 2 . 74 . 7 2 96. 3 85-90 3. 500 . 030 . 2 . 60 . 5 0 97. 0 90-95 3. 700 . 025 . 2 . 50 . 5 0 95-100 3. 900 . OZ) . 2 . 40 . 4 0 . 100-105 4. 100 . 017 . 2 . 34 . 3 0 . 105-110 4. 300 . 014 . 2 . 28 . 3 0 . 110-115 4.500 .011 .2 .22 .2 1 . Approaches 98. 88% 87. 7 89 Approaches 0 Approaches Approaches 100' 00% 100. 00% 89 1 ( ) Bracket used to capture the length of time interval between signals. Example: Elapsed time of 77 minutes between signals falls into the 75-80 bracket (a fixed 5 minute wide At bar) for histogram purposes. 53 .Hd030u005 .m> Hmsuomnnumchfi H9335 8069mm 0%. uOH 3.9535 Hmzwwm mo Cofldnwnumwfl A: ousmwm + 1 v m N _ o. m. w. N. a . . J . . . . J m.. 330m S .073 Homdcmfi new 3.5...me no 003 £03533an .. . , . u , , .w #u‘. n . d - coo rag com «up cow rm» oo¢.omn Con ova 33:.aon nom.1c~ :m:.~N- p. .ms . u.r. ta..u —ooooooooo~ouo.oou.o~ooooooooo—coo...uo ~o...oo.o.u_.oo-ooooo_-uooooo-o‘oncogene-won... ...h..o...o... a°.n . , . bllcaes.-wfl.lllt 3.3 . n :30: u NtP.:B¢S.nnr)|1 ao.o~ . o ao.on . . o .o.o- a o o no.9“ » o u o . o . Z . €17.08. . 22.138323. 5333.6 33.5.6 accuses-i o A .o.o~ . a o a ao.oo H a; . «aaxa u “aux- . awaaanamnxx .o.oo n «auauuuauauunuuauauaaawannaaaaauaaauafiaaaaaannnnnmxxxxuuxuxwx .xxx-xxxxxxxxxx-xxxxxxxxxxxxxxxxxxxuxxxxa‘xxxuxx444xx ... .«an H 0009. xxndd . cacao-ah v» 414 oocecoaoncc xxxna .o.oom .pMuw-nu4m scrcwrz_ox \ \on~.mm . sue 2.» amps... u 2.0: oaoo.coo - uaz Opportunistic Craftsman The manager (O-E) (C-E) Distribution of entrepreneurial types. Figure 13. V. RESULTS OF TIME-USE STUDY 1 "Zeit ist Geld" Since the turn of the century ushered in the beginning of the "Scientific Management Movement, " industry has spent considerable time, talent, and money producing analytical studies about its work- ers. Countless hours have been devoted to measuring, interviewing and studying workers' activities in stop-watch dimensions. This re- search was prompted by the desire for increased efficiency. Yet. as Brisley (6) points out, the executive. . . or manager. . .or entrepreneur. . . has escaped this close-up analyzing. This is not surprising when one studies man's attitudes relative to the times. Until the depression, the division of labor was clear-cut. Rank (especially ownership) had its privileges, and labor was just beginning to flex its muscles. At the same time. a gradual shift was taking place in management's view of the worker. Management had. tradi— tionally treated the worker as the mechanistic means to achieve the the production goals; it now began to recognize him as a person with many complex needs. A major turning point in industrial relations 1"Time is Money. " 75 76 could well have been the famous Hawthorne Study1 (34) at Western Electric Company in the late 1920's. During the 1930's contributors to the business fields from the social science disciplines were gathering under the "human relations" umbrella. The war and post war years saw periods of increased technology. a rapidly expanding economy, and the courting of the worker by man- agement. At the same time management was also learning more and more about automation. Eventually the inevitable happened: in the name of efficiency and costs, entire operations as well as individual operators were replaced by machines. It seemed that while the intent had been to take the engineerirg out of "human engineering, " the human had been engineered out instead! Inevitably. the cost-cutting or examining finger had to point to the managers themselves. Suddenly, the boring of six holes instead of five, with its $100, 000 annual savings in labor and material. be- came less serious than a managerial fl pa in decision making. which could cost the business many times that amount. Thus as industry becomes more highly sophisticated, much more must be known about managerial behavior and effectiveness. 1This study showed that in addition to the formal organizational structure. informal networks that have great influence also exist. It dealt with motivation, productivity and quality of work as they relate to social relations among workers and between workers and their superiors. 77 This knowledge is as basic to the success of small managers as it is to those in big business. The time-use data collected with the Random Interval Timer recorded a total of 2215 observations. Less than 300 of these occurred after 6:00 p. m. and will not be included in the tabulations. These night time observations were irregular because the managers turned their devices off at different times toward the end of the day. Although these observations were too incomplete for calculating per- centages of total time, they do illustrate that the men continued to be engaged in business activities during the evening hours. Some man- agers returned to the plants to do maintenance, problem solving and routine work. The actual number of night time observations for each activity can be found in Appendix E. Table E2. Hypotheses Tested in This Study The hypotheses stated were formulated in the planning stages of this research based upon knowledge of the industry. 1. Hypothesis: The Opportunistic-Entrepreneur spends more time in the decision making process than the Craftsman- Entrepreneur. 2. Hypothesis: The Opportunistic-Entrepreneur spends more time in the communication processes than the Craftsman- Entrepreneur. 3. Hypothesis: The Craftsman-Entrepreneur engages in more 78 manual labor in his firm than the Opportunistic-Entrepreneur. 4. Hypothesis: The Opportunistic-Entrepreneur participates in more training programs than the Craftsman-Entrepreneur. Smith's (38) method for identifying the Craftsman-Entrepreneur and the Opportunistic-Entrepreneur was basic to proceeding with this time-use research. If these hypotheses could be verified by the time- use results of the sawmill managers, then the work sampling approach might provide a useful tool in identifying these managerial sub-types without the need for subjective analysis. Percent Time-Use This portion of the time-use data results analyzes the percent— age occupied by each activity of the total time sampled. The data is assembled in an order that progresses from a profile of the entire group to an examination of the sub-group comparisons in order to accept or reject the previous hypotheses. All eleven manager 3 One of the purposes of this research was to view the managers of small sawmills as a characteristic group of the Forest Products industry. Figure 14 illustrates the percentage of the total daytime observations devoted to 13 activities, as listed on the Data Card. In Figure 14, the category OTHER on the Data Card was divided into Manual Work, Home and Personal, and Meetings, to more clearly 79 1 Activities Percent of total time 0 10 20 30 ROUTINE mRK V/ l I l// I ‘ I .......... ////I. 2:23:33. //////1. .4. m... ///////. GIVING 6. 64 INSTRUCTIONS A HOME, PERSONAL2 // 5. 39 A CORRESPONDENCE //l 3- 84 :0 8 RECEIVING A 1 81 PROBLEM A ' JOB INTERVIEWS .98 A 2 MEETINGS .78 TRAINING .10 1 Based on 1929 daytime observations. 2 Part of category "Other. " Figure 14. Activity profile of all 11 sawmill managers. 80 define the manage rs' activities. The managers spent the largest segment of time. 30. 33% of the total daytime hours, doing ROUTINE WORK. This area had been divided into four sub-activities. Examples of the information gathered show that managers used this Routine Work time in the following ways. 1. Consultation with subordinates; usually guiding someone through a routine assignment such as instructing office help in form-keeping or a yard man in placing finished lumber. 2. Consultation with the outside; could be talking to an equip- ment salesman or a potential log supplier. 3. Scaling logs or grading lumber; a technical function often performed by the manager. 4. Office work; calculating log scale. lumber footages. accounts. payrolls, etc. , indicating that the manager func- tioned as an accountant, or salesman, or personnel manager. MANUAL WORK accounted for 10. 88% of the managers' time. Since this was a managerial study. a percentage this high may sur- prise those unfamiliar with small sawmills. The time devoted to the problem solving or decision making pro- cess (RECEIVING A PROBLEM. GETTING INFORMATION ABOUT A PROBLEM. DECIDING A COURSE OF ACTION. and GIVING INS'I'RUC- TIONS) accounted for 27.47% of the managers' total time. This in- cluded both oral and written aspects of the problem solving process. These were listed individually in order to identify and analyze the 81 time spent in each step. The discovery that only . 10% of the managers' time was spent in TRAINING shows that this was an almost non-existent activity. High Opportunistic-Entrepreneur vs. High Craftsman— Entrepreneur In Figure 13, Chapter IV, a cluster of three managers was polarized at each end of the continuum. These two clusters repre- sented the more "ideal" entrepreneurial types, the High Craftsman- Entrepreneur and the High Opportunistic-Entrepreneur. These sub- jectively divided clusters were tested statistically to see if there was truly a difference between the two pure managerial types on the basis of their use of time. Figure 15 illustrates the percentage of time spent in the 11 major activities noted on the Data Card by the High O-E and High C-E types of managers. The decision making sequence was noticeably different between the two pure types. The O-E's spent 33. 26% of their time in problem solving activities as contrasted to 15. 48% spent by the High C-E's. This study did not provide information to show w_h_y the O-E's spent 114% more time making decisions. Perhaps more problems were brought to them. or they were more inclined to seek out the problems. Or perhaps they had a greater desire to deal with these things. But the fact remains that the Opportunistic-Entrepreneur did spend con- siderably more time receiving, getting information. deciding a course 82 1 2 Three High O-E type managers (2, 8, 10) Three High C-E type managers (4, 6, 7) Percent of time 35 30 25 20 1.5 10 S 0% S 10 1 5 20 25 30 35 1 I L .L 1: la l AL I L l J 1 l I TRAINING 0 . 26 RECEIVING A PROBLEM 2.16 /1 GETTING INFO so... . m... .. .. //////-81s DECIDING ACTION 11 70 (PROBLEM) //m 3.14 m” "/////// //// JOB INTERVIEWS 9: 1.56 V/ CORRESPONDENCE 5.92 TRAVEL 12 9° ////// W... m '///. $35.: Work, etc.) '0‘85 m . 10. 31 1 Based upon 511 daytime Observations and pooled averages. 2 Based upon 521 daytime observatiom and pooled averages. Figure 15. Percent of time spent in 11 major activities by High O-E's vs. High C-E's 83 of action. and giving instructions about problems. The overall proportion of time expanded in ROUTINE WORK was almost identical for the two "ideal" types. This category, which used 27% of the managers' day. illustrated the variety of job roles per- formed by the small mill manager in both sub-groups. The larger CORRESPONDENCE time of the Opportunistic- Entrepreneur coincides with the hypothetical view that the flow of communications should be increased with this type of person. The large percentage of time spent in OTHER by the High Craftsman-Entrepreneurs represents the most striking single time- use difference. This category included manual work, home time and meetings. Because these were not listed individually, the Event statements on the Data Card provided the means for sorting out the numbers of observations in these sub-categories. A more detailed analysis of all observations for the High O-E's and the High C-E's is shown in Figure 16. The actual number of observations when used to compare the sub-groups, reveal that: l. The Oppo rtunistic-Entrepreneur spent more time communi- cating in the decision making process. When GETTING INFORMATION ABOUT A PROBLEM he turned to outside sources and subordinates more often. Consultation with subordinates was also evident when DECIDING A COURSE OF ACTION. and he was more verbal when GIVING INSTRUC- TIONS. 84 Three High O-E type managers (2, 8, 10) Three High C-E type managers (4, 6, 7) 1 Number of observations 30 Irlrrrlrl TRAINING Formal within Formal outside RECEIVING A PROBLEM Mail Phone In person GETTING INFO ABOUT A PROBLEM Subordinate Outside source Technical Reading Other (family, corp. officers) DECIDING ACTION (PROBLEM) Observation time With subordinates Think time Other GIVING INSTRUCTIONS 35 201510 5 0 5101520 I FT ITITIIT' 0 1 19 17 14 10 14 Oral Written ROUTINE WORK 34 14 Consultation with sub. __ L Consultation with outsidL Scaling, impection, etc. Office work, figuring, etc. JOB INTERVIEWS CORRESPONDENCE Written or dictation Reading TRAVEL M Co. business L To 8 from home 7 BREAKTIME Coffee, etc. Meals OTHER Home, family, personal Manual work in plant Meeting 51 63 L 31. .1 12 10 10 49 39 . 12 1 O-E NO. has been adjusted to C-E's 521 from 511 for the purpose of this chart only. Figure 16. Actual number of observations recorded by High O-E's vs. High C-E' s. 85 2. The Opportunistic-Entrepreneur spent more time in COR- RESPONDENCE. and he used the largest part of this time in reading the correspondence rather than in writing or dictat- ing. 3. While the total time spent in ROUTINE WORK was nearly equal for both groups. the Opportunistic-Entrepreneur spent more time communicating in this category and less time in those activities that resembled other work roles. Apparently scaling, grading and office work were activities that the O-E would rather delegate to others. 4. Company business accounted for almost the entire TRAVEL category for both groups and a great deal of this time was devoted to looking for raw material. 5. The High Craftsman-Entrepreneur spent a larger percentage of time in coffee breaks and meals. 6. The High Craftsman-Entrepreneur spent more time doing manual work around the plant. Average time allocated to activities In Figures 17 and 18, the average 10 hour (600 minute) day shows the average total time spent in each activity. This was deter- mined by multiplying the 600 minute day by the pooled percentage of time each group spent in each activity. The results for the three High Opportunistic-Entrepreneurs are shown in Figure 17 and for the three 86 HOME, src. .. 10.92 min. gTRAmINC .. 0.00 min. MEETINGS - 14. 76 min. ' gRECEIVING A PROBLEM - 12. 96 min. GETTING INFO -73.80m1n. MANUAL LAKDR - 39.42 min. MEALS, BREAKS- 52.80 in. m N . 0 . DECIDING ACTION 70. 20 min. TRAVEL- 77. 40 min. crvmc INSTRUCTIONS- 42. 60 min CORRESPONDENC571 35. 52 min. ' JOB INTERVIEWS .. (\ z ROUTINE WORK - 164. 16 min. 5. 34 min. * Figure 17. Total number of minutes per day1 devoted to key activities by three High O-E managers. 1 Based upon average 10 hour day or 600 minutes. Divide by 10 if per hour estimate is desired. TRAINING - 1.56 min. HOME, ETC. - 43. 74 min. RECEIVING A PROBLEM — 1.08 min. EETTING INFO - 42. 59 min. J DECIDING ACTION- 30. 06 min. GIVING INSTRUCTIONS- 18. 84 min. MANUAL LABOR - 136. 86 min. O O O . 990.3099, .9.0 o e 92-” o o o ’9 O .3. 0:0 0.9 90 9%... . 0.9.00 0%. r ‘o .0 ,0. o o 0 9:0 ’0 \ O 0 ROUTINE WORK— 165. 12 min. JOB INTERVIEWS - 9. 36 min. CORRESPONDENCE - 7. 44 min. MEALS, BREAKS - 81. 12 min. TRAVEL - 61. 86 min f 1 Figure 18. Total number of minutes per day devoted to key activities by three High C-E managers. 1Based ,upon average 10 hour day or 600 minutes. Divide by 10 if per hour estimate is desired. 88 High Craftsman-Entrepreneurs in Figure 18. By shading and cross- hatching, the differences between the sub-types is shown more dra- matically. The most noticeable differences are the decision making components and manual labor. High Opportunistic-Entrepreneurs vs. all Craftsman- Entrepreneurs As shown in Figure 13, Chapter IV, five managers were clus- tered near the center. from (0) to (-3), on the continuum. These managers exhibited characteristics from both "ideal" types. although their final scores placed them on the Craftsman side. To compare the High Opportunistic-Entrepreneurs with all of the Craftsman- Entrepreneurs, the observations for the middle Cluster of five were added to those of the High C-E group. When all of the C-E observa- tions are grouped together they tend to move the entire C-E group toward the O-E group, making the differences less obvious (Figure 19). Statistics used in comparisons Since anything can be proved by statistics, the researcher is faced with a dilemma. He must use the collected data without manipu- lation. Sample sizes were uneven in this study making it especially necessary to select methods of handling the data that would provide a fair analysis. The activity profile (Figure 14) of all of the 11 managers is based upon the number of daytime observations recorded by the 89 Three O-E type managers (2, 8, 10)1 vs. All C-E type managers . (1, 3 ,,4,5 6, 7, 9, 11)2 Percent of time 35 30 25 20 15 10 5 0 5 10 15 20 25 30 35 L i l m a 1 _L TRAINING .14 RECEIVING A PROBLEM 2. 16 1 2.08 V ammo INrO ABOUT // A PROBLEM 12.30 10. 21 DECIDING ACTION (PROBLEM, 11.70 ///// 7. 31 GIVING INSTRUCTIONS 7.10 / 6.95 ROUTINE l-lll my. 36 / I'll, IOBINTERVIr-zws .89 1 1.04 m m: MEALS AND BREAKS .. // $3.33.... s .. /// 2.92 18.60 1 Based upon 511 daytime observatiom and pooled averages. 2 Based upon 1418 daytime observation and pooled averages. Figure 19. Percent of time spent in 11 major activities by O-E vs. all C-E managers. 90 participants. The percentage of the day that each manager was en- gaged in each activity was the ratio of the number of observations of that activity to the total number of observations. Work sampling is represented in the standard "ratio-delay" form in this application. In this statistical group comparison the role of the individual manager was secondary to the entire group. The time-use percentages for each individual manager can be found in Appendix E, Table E3. The problem arose when the Opportunistic-Entrepreneur was compared statistically with the Craftsman-Entrepreneur. The num- ber of observations for each manager were not the same. A manager with more observations would influence those with less, and thereby change the averages of the group in which his data was pooled. Figures 15 and 19 reflect the pooled averages and help correct this problem of uneven sample size. Table 10 compares the ratio of actual observations with the pooled averages of each sub-group. The figures do not change a great deal when the two methods of calculating the time-use percentages are compared. A procedure for statistically testing whether two samples. dif- fering in size. come from different populations is described by Snedecor (37, Section 4. 9). His variation of the Student's t test went one step further than pooling the averages. Using this method. each day's percentages were tabulated by computer for each man. The expression used to calculate the t values was: 91 509308 mo non—59a ¢ nomads-n 20m 0w80>u Fat-«80W £835,830 «0 200.55 H38 IuI £300 .200 3233030 inn-00 .«O con-:52 u 00.0.. 00.2: 00.00 00.00“ 00.00 00.22 w0.00 coded o0 H.068 waif 00.: abma Nm.0~ o~.0N omém mwéfi 2.4: A000 40:92): mmmHO mm.o.. Nm.o~ mN.: 0mg: Nm.m~ ANNA 0mm 5N mMmHHZH mOh 0m.om mm.om Nmém Nmém Nm.>N océN oméN HofiN MMO? MZHHDOm 006 No.0 m0.o omb Al .m mwN OH .0 mmO mZOHHODMHmZH 07:20 HmN mm.w Hm.” 00.0 Ho.m wim 00.: 00.: ZOHHO< UZHQHOHQ 322 50.2 NA: 2.0 :4. mmb omNH NmNH ZOHHEMOMZH DZHBHMO ofiN $4 woN 004 Nd. 0H. ofiN o0; EBmOMQHMOHM OH. 0“. 0H. 0H. 0N. 0H. I I III UZHZH< .N0mfi0 .0>< 3000.0 .0>< @0030 35304 p0~oon~ H030< p0~oonm 330.4 p0~oom H0304, p0~oom 3300. Ho .«o .«o no ofism Oflsm ofiom 030m mammmcmz 0n0mmcm§ mu0monm§ mu0wmndz :4 m0 :4 m0 swam M mo :05 m $33000 £000 MOM N H momma0>m @3000 .m> @0508 >30OIOENN mo GOmENQEOU . OH 03.08 92 2 Z .83.. + 2. l 1 n2 2 t:caIl : and ttable : s 2 s 2 -1— + _..2_ n1 2 where: ;l';2 = the means of samples 1 and 2 Z 2 , 31.32 = the variance of samples 1 and 2 n1. n2 = the number of observations in samples 1 and 2 t1. t2 = the table value of t at (n-l) degrees of freedom for samples 1 and Z. Snedecor's (37, p. c)8) modification of the Student's t test was to be used for testing. "Two samples, differing in size, from populations with different standard deviations (01 a! 02), together with the test of HO : “I = p2 . . . .", meaning that both samples come from the same population. Results for each activity are shown in Table 11. The calculated t and ttable values can be found in Appendix F. The statistical interpretation from Table 11 is that in the major- ity of their activities the High Opportunistic-Entrepreneur and High Craftsman-Entrepreneur represent two distinct and. separate sub- ty es (H is re'ected or p a5 p ). Thus the modified Student's t 0 test helped add credibility to the already discussed visual differences of Figure 15. 93 Table 11. Tests for significance by activity. t Calculated Three High O-E vs. Three High O-E vs. Three High C-E Type All C-E I‘ype Activities Managers Managers TRAINING -- N. S. -- N. S. RECEIVING A PROBLEM Z. 5791 * —- N. S. GETTING INFO 3.4962 ** -- N.S. DECIDING ACTION 3. 3737 ** 2.5678 * GIVING INSTRUCTIONS 3.1431 ** -- N. S. ROUTINE WORK -- N.S. -- N.S. JOB INTERVIEWS -— N.S. -- N.S. CORRESPONDENCE 3. 3354 ** 2. 1134 * TRAVEL 2.1650 * 3. 2536 ** MEALS and BREAKTIME - - N. S. -— N. S. OTHER 7.9296 ** 4. 3343 ** N.S. = not significant. * Probably significant—-5% level. ** Highly significant- -1% level. When all the Craftsman-Entrepreneurs are included. the "ideal" types are less distinguishable and it became difficult to test for sub- types. The Results and the Hypotheses The hypotheses stated early in the chapter can now be examined in terms of the time-use results. Hypothesis: The Opportunistic~Entrepreneur spends more time in the decision making process than the Craftsman-Entrepreneur. The percentages shown in Figure 15 indicate that the three High O-E's spent 114% more time in the decision making. or problem 94 solving. process than did the three High C-E's. A further breakdown. comparing the two pure types of managers according to each event shows that the O-E's spent 12 times as much time RECEIVING A PROBLEM. almost twice as much time GETTING INFORMATION ABOUT A PROBLEM. and over twice as much time DECIDING AN ACTION and GIVING INSTRUCTIONS ABOUT A PROBLEM. Further evidence is found. in Figure 16. comparing the actual number of obser- vations. The problem solving group of activities shown in the shaded areas of Figure 17 and 18 provide a clear visual picture of these dif- ferences based on the 10 hour day. In view of this evidence. the hypothesis is accepted when considering the two pure managerial types. Hypothesis: The Opportunistic-Entrepreneur spends more time in the communication process than the Craftsman-Entrepreneur. When observing the communication process strictly on the basis of numbers of observations. a definite contrast can be seen in Figure 16. In all of the categories that would involve the communication process (with the exception of JOB INTERVIEWS and TRAINING). the number of observations for the High O-E's is greater than those for the High C-E's. These categories included the four factions of the problem solving process. some portions of ROUTINE WORK. and CORRESPONDENCE. While a more complete analysis of the commu- nication process will be made in Chapter VI. the hypothesis is accepted when considering the two pure managerial types. 95 Hypothesis: The Craftsman-Entrepreneur engages in more manual labor in his firm than the Opportunistic-Entrepreneur. Manual work was not listed as a major category on the Data Card but was one of the activities recorded as a part of OTHER. Fig- ure 16 shows the finer breakdown of the categories and indicates that the High C-E spent 309% more time doing manual work in his firm. The crosshatched areas in Figures 17 and 18 give a good visual pic- ture of the manual labor comparison for the two groups. Based on these findings, the hypothesis is accepted when comparing the two pure manage rial types. Hypothesis: The Opportunistic-Entrepreneur participates in more training programs than the Craftsman-Entrepreneur. Out of the total of 2215 observations, there were only two events recorded for TRAINING. The hypothesis is therefore rejected. When comparing the Opportunistic-Entrepreneur with the total group of eight Craftsman-Entrepreneurs. it becomes difficult to draw significant conclusions relative to the hypotheses. Figure 19 indicates a slightly higher percentage for the O-E's in the problem solving area than for all of the C-E's. However. this was not sufficient for accept- ing the entire decision making hypothesis. In Table 11, the statistical significance of OTHER (which is pre- dominantly Manual Work) indicates that if the O-E's were compared, to the total C-E group, the hypothesis concerning manual work would still be accepted . 96 The existence of a middle group of managers. not consistantly aligned with either of the pure types. cannot be ignored. Work sam- pling alone could separate the polar types Craftsman—Entrepreneur and Opportunistic-Entrepreneur. but would not provide a reliable approach for separation when considering this clearly existing middle group. Therefore. if a definition of managerial types is needed. some form of subjective evaluation will continue to be necessary until further work sampling research has been done. Average Duration of Each Event The average time spent on each occurrence of an activity by the managers was determined by an entirely new concept. Several intri- cate systems have been used successfully to establish job length (36) or time frames (26); each requires a high rate of sampling and/or a low number of tasks to record. The data collected in this study showed that the managers were involved in 27 types of activities. mak- ing the existing systems unusable. To test the duration of an event. each mill manager was asked to record the amount of time he had spent in the activity caught by the random signal. This was not to be the total time spent, but the amount of time involved in that event prior to the alarm. As an ex- ample. if he was talking on the telephone when the Random Interval Timer sounded. he not only recorded the clock time of the signal. but the time in minutes he had already been talking on the telephone. If 97 he had been receiving a complaint (RECEIVING A PROBLEM) for three minutes, it was considered the "half life" for that activity. The three minutes were then doubled to estimate a six minute activity length. The average duration of time spent performing each type of activity could then be calculated. The author felt that the long and short times would average out if a substantial sample was taken. No statistical tests we re run on this data. To arrive at the averages shown in Figures 20 and 21, the total of all durations re- corded for each activity was divided by the number of observations for that activity. The amount of time spent when the event occurred was the important factor: not how many managers performed the activity. but the average time it took him if he did. This way the non- participants in any particular activity could not affect the averages for those who did participate. Figure 2.0 compares the Opportunistic-Entrepreneur with both groups of Craftsman-Entrepreneurs. An abundance of observations by one operator could have influenced the duration averages. The calculated average remains a truthful representation of the facts for each group of managers because non-participants we re not included. What catches the eye of the analyzer is the longer times spent per activity by the Opportunistic-Entrepreneurs with the exceptions of TRAINING. ROUTINE WORK and JOB INTERVIEWS. Does the O—E take more time. is he provided more time by his organization. or has he delegated responsibility well enough to provide himself more time ? 98 Three High O-E managers (2, 8, 10) vs. I Two1 High C-E managers (4, 7) l r 1 ' A11 C-E managers (1, 3, 4, 5, 7, 9, 11) B 69 59 4p 30 29 19 o 10 20 30 40 so 60 70 I f r r V I m wt r r I v TRAINING 0.00 20'00 RECEIVING A PROBLEM 21.40 0'00 m 28.67 GETTING INFO ABOUT 32 75 11.68 A PROBLEM ' ”ll/A 27.57 . 22 DECIDING ACTION 37.60 V] 15 l 30 25 6.50 ‘ . GIVING INSTRUCTIONS 25. 27 m 40 26 ROUTINE WORK 27.25 9'2: 7s 3. 33 JOB INTERVIEWS 22.50 7 48 29 CORRESPONDENCE I 37.33 , ‘ '3' , m 34- so g; 23.13 TRAVEL 77.16 VII/A 24.65 7r 29.40 MFALS a BREAKS I 33.33 W4 32_ 48 l 37.88 OTHER 54. 73 7 43' 45 1 Manager 6 did not recoxd this portion-data not included. Figure 20. Average duration of each activity in minutes, comparing High O-E's vs. High C-E's and all C-E's. Activities Average time in minutes 40 //8///// [m TRAVEL Em“ w 12:... m... //////////////////////////A 4° W in.» NW W /////////////////////// 5.2m: “35° W “//////////////////m W m////////////w 100 In contrast. does the shorter duration of events for the High C-E indi- cate a greater degree of efficiency, or is he racing from ”brush fire" to "brush fire"? The research is not conclusive. but Offers an inter- esting avenue for further study as the efficiency of managers gains interest. When all of the Craftsman-Entrepreneurs were grouped. together. their resulting duration of events was closer to that of the Opportunistic-Entrepreneurs. In many activities the middle Cluster of five managers had slightly longer average times which offset the shorter averages of the three High C-E's. Figure 21 is a profile for the entire group of managers and represents 1744 observation-duration times. All but one Of the deci- sion making components fall into the lower portion of the Chart. indi- cating shorter times spent in these events. At the bottom of the chart is TRAINING. the most neglected activity for all of the 11 managers. While focusing upon the duration of an event. the frequency of each activity must not be neglected. A manager spends an average of 32. 02 minutes per ROUTINE WORK task. but he also performs this task more often than any of the other activities (Figure 14). In the same way. the 20 minute duration of TRAINING must be kept in per- spective by observing the rare occurrence of the event. lOl Comparison With Existing Studies A time-use comparison Of the sawmill managers with other groups of executives was considered worthwhile. even though very little other research was available. Seven studies from four sources are compiled in Table 12 according to the percentage of time spent in various activities. The four columns on the left side of the table are the author's attempt to classify similar categories from these studies. Some broad interpre- tations were necessary. since term definitions differed from one study to another. The percentages of the three High Opportunistic- Entrepreneurs are shown in addition to the totals for the 11 mill man- agers. The data collected in this study relates them more closely to big business than are other managers in the small sawmill group. The decision making component is the single area that allows some measure Of comparison. However. the variance in category definitions makes it impossible to form definite conclusions. Refer- ring to the three studies in the Social Science Approaches to Business Behavior. Dubin (17) states. If these studies are at all representative of what executives do. it would seem that making decisions. which is often considered their cardinal function. occupies a remarkably small share of their total working time. This statement would also seem to‘apply to the Brisley and sawmill manager studies. in view Of their percentages of time in decision making categorie s. .803805 85:88.0 8. 800800 .8800 8882 88% E .8 d .9 028.... Am. 8:80 8.. 8.. a .83 258 8a EBB... 820 888.. 5 d .3 380 E .8 s .8 813. 8a 3830 3 8-8-58 .......................................................................... .0.- figw a e iama .................................................................... a .mB-mwwwfi ................................................ s... ............................................................... .- 88.-m fink-1.8mm”--- @- .m.-I-momma-58.5---be-mIiB. MIN-I..-- -- -II-I~.0.. .................... Mum-.mp- --- :wmwmmommm- ............ .. ww$.-Iwmwwm§.m.mmm mam-5m ...... sum .................................... 0.8.8.2852 8...: as... ............ 23.280808..an838 8 S 8 a 8.2 3803 .oo 880 80:3 shim...“ .......................................................................... 05 mm ................ 8 .m 8 .m 8 N 88.3 w 8885 8885 II._---II--I.-I-----.. ......................................................... m-ommmwba-mmmmbw ................ s ...... m... ...... 80 .......................................................... MsssmME-mfi---8.8.8.882 mm.m----mmI.m----B-B ....................................................... mmmmmmme...fim.me 880880 3.8.0.8385... ................................ 0.-.»---008555.8:»---e..m.--s..m..:-s..m-. ................... 8% 88.-2.2.088 ..................... m. mac-£08 ............... 8.: mm .m 2 .fi 8 .o 8.3 8.8 8.8 8.8 88 20:04.. 8:8 8 8:80 .833qu 88.4.0803 3:88 ....... ”.888 ---------- mag-:55Wfime 8 2 8.0 8 o 8 E 8.2 8.2 8 2 «N a 8.8 33? 808500 8339.... 32.803 803380 w 8835. .333. IIIIIIII immufinaou 8 a 8 3 20895 08.88 8 .2 0...: 8.2 8 .8 8 .8 8.2 8.8 8 .8 80 22...ng 88582 883 0200.50 580 8880 w 33 .15. 8: dun: .20 8825 .30 .mo .82 mm: .mm-Z 8838 E .335 E E 8:2.” MIC 52 30.500 mamas-92 0.80 E82 .03.— ._um “caesium mfiafism 83930 E .33. E a; .283 8v .8...— i: £- 88.88... m z 038.... 888 8 E 5 883m 328 3. E 8338 888 £- 838 8:588 3888.. Go 8359.883 #38080 258. 88800 8300295. oogmow Humoom \03950 60:03» 0110fi>30u >9 000% 083 035038 .«0 300.80 .3 030.0 103 Brisley's study of Detroit Executives can be compared more closely with the mill managers because the category divisions were more similar. Brisley's study did not have any activity corresponding to Routine or Manual Work. The complete omission of training in all of the studies, except the .10% for the sawmill managers, is interesting. When does the manager keep up on the new tools of his trade ? Another obstacle to a realistic comparison was the two different types of executives. Those in the comparison studies were big business executives, or more like them. No time-use research could be found concerning entrepreneurial executives. I03» PART III. MORE DIFFERENCES: COMMUNICATION ORGANIZATION Communication 9 VII. Organization VI. VI. TYPE OF MANAGER: HIS COMMUNICATION AND ORGANIZATION PATTERNS Integration of Communication and Organization Theory Two students, one studying Organization Theory and the other studying interpersonal communication theory, would eventually find themselves searching and. reading the same literature. How man organizes himself for accomplishing tasks relies so heavily upon how he communicates that the areas become inseparable. The traditional flaproach The role of communications has been treated very differently by the traditional and modern theorists. Organization Theory is not old; it started in the early 1900's with contributions from Taylor, Gannt. and the Gilbreths. In the early years it was called Scientific Manage- ment and later writers have used the term Administrative Design Theory. Originally these theories were built upon the concept of a clear division of labor. From this point of view, maximum efficiency or productivity could be attained only when a division of work was achieved at all levels of the organization. Men were looked upon as adjuncts to machines and as occupants of job descriptions in 104 105 structured organizations. Management was the rational element (according to management). enjoyed rigid control, and. communicated from the top down with little feedback. Evolution of modern theory Meaningful consideration of the patterns of human relationships began to change the emphasis in the 1930's. Studies of communication networks also gained attention. The discovery of feedback channels- or bottom-to-top communications, demonstrated the sensitivity that was developing to this vital part of organizational life. After World War II, modern Organization Theory began to evolve. One of its main concepts is the importance of each member to the formal and informal structure of the organization, and his con- tribution to these structures. The new decision making processes and. management-by-objectives generated interest from both researchers and practitioners. The organization was viewed as an ”open system" rather than a "closed system. " Finalbr some recognition Today the term communication is overworked, underpracticed. and confusing, even to those trained to teach its wide spectrum of intrepretation. While the management scientists were enlarging their theories to include communications, the social scientists were direct- ly linking communications to human behavior. A few of the outstand- ing organization theorists show a wide range of views that run from 106 information acquisition systems, to the effect of networks and their communication flow, to change agents (and why they change opinions), and to interpersonal communications (and behavior change). March and Simon (25) list channel usage, content of the "in group" communication, efficiency of communication, and instruction of communication among a series of variables in their theory. Likert (Z4) treats communication as an intervening variable and introduces the idea of linking key people and their groups together in organization networks. McGregor (Z9) and Herzberg (19) leave no doubt that as we pro- gress toward participative management, the value of being a good communicator is an essential part of a total managerial philosophy. Communication and the Sawmill Manager The need to know about the sawmill manager's communication pattern stems from the concern for the high mortality rate in the small sawmill business. By virtue of his position, it was assumed that the manager was at the apex of the communication network in his firm; that he was in the position to affect those who work for and with him. Knowing about his communication habits should show one dimension of his effective- ness as a manager. But knowing who he talked to and listened to would provide another, even more important, kind of information: what avenues are 107 available for reaching the manager with any form of help or informa- tion? In the rapidly changing business and industrial environment, new ideas and techniques are constantly appearing. These come in the form of education and training, product innovations, quality control improvements, and technical advances in equipment and utilization. These advances are virtually useless unless they are received. The combined time-use, organization, and communication information could provide knowledge about available channels for reaching the manager. The communication data was collected. with the Random Interval Timer. If there was some form of communication going on at the time of the signal, this was recorded on the Data Card. Several important aspects of the communication process could not be captured by a form of data collection that observed only the dyadic communication network of the manager. It could. not be shown if the verbal communications involved a one-way or two-way system. Leavitt (23) states that while the one-way system is faster, looks orderly, tends to hide the sender's mistakes, and protects his power, the two-way system is more accurate. It is also more demanding of the participants. Thus it was impossible to measure the effectiveness of mes- sages in these interpersonal communications. The importance of the listener or receiver, and the percentage of listening or receiving 108 required for a manager to be effective, cannot be overlooked. Still the volume of oral communication flowing to and from the manager. and the channels he used inside and outside the organization, provided an important view of his communication pattern. Total time spent in communication Berlo (4, p. l) cites that ". . . the average American spends about 70% of his active hours communicating verbally--listening, speaking, reading and writing--in that order. " Dubin (l7) summarized two industrial studies conducted on foremen of two large companies. The figures for time spent in communication by these foremen ranged from 46. 6% to 57.3%, depending on whether the observer was counting actual conversation or all ”interaction” contacts. Brisley (6) re- corded that the managers in his Detroit study spent 80% of their time in oral communication. Table 13 shows that five of the sawmill managers fall in the 40% to 60% range. The average for the total group of 11 managers was 35.46%. The average communication time for the three High Opportunistic-Entrepreneurs was 48. 53% and falls at the lower edge of the industrial range when compared with Dubin's findings. The three High Craftsman-Entrepreneurs spent only 19- 70% of their time in communication. The time-use and interview information provided a possible explanation for the difference in communication patterns for the two 109 groups of managers. In general, the Craftsman—Entrepreneurs operated smaller firms which demanded. that they fill more job roles, such as manual labor and maintenance. This would leave them less time for typical managerial tasks. In contrast, the Opportunistic- Entrepreneurs spent more time in their larger firms doing managerial type activities, and therefore spent more time in some form of com- munication. The fact that the High O-E's engaged in some form of communication 2.4 times as often as the High C-E's further supports the hypothesis stated in Chapter V. Table 13. Percent time spent in communication. Pe rcent of Communication Percent of Total Obser. Observations Activities Involving Manager is Manager is Group Communication Originator Receiver All Managers 35.46% 81.29% 18.71% 3 High O-E type Mgrs. 48. 53% 82.47% 17.43% 3 High C-E type Mgrs. 19. 70% 88.35% 11.65% Manager 1 36.73 78.70 21.30 2 40.88 76.92 23.08 3 55. 26 66.67 33.33 4 15.38 81.25 18.75 5 20.47 88.46 11.54 6 22.70 88.10 11.90 7 22.66 96. 55 3.45 8 54.36 81.48 18. 52 9 31.31 91.94 8.06 10 50.25 87.25 12.75 11 45.24 75.44 24.56 110 Dire ction of flow A total number of 1929 observations were recorded. by the Random Interval Timer for the 11 mill managers; 684 of these caught the managers in some form of communication. These 684 observa- tions provide the base for examining the communication percentages. When a communication activity was recorded, each manager was asked to state on the Data Card whether he had originated the con- tact or acted as the receiver. Table 13 shows that the managers acted as the sender, or originator, in more than 80% of the communi- cations. The actual tabular record is shown in Table 14. Illustra- tions of the communication flow are provided in Figures 22, 23, and 24. Comparable research about managers reported by Dubin (17) showed that when managers initiated contacts inside the firm, the fig- ures ranged from 57% to 74% of his communication time. The vari- ability depended upon which subordinate group was involved. Figure 22 shows that the total group of sawmill managers initiated communi- cation inside the firm 48% of the time. Manager initiated communi- cations were directed outside the firm 33% of the time. Only 19% of all the managers' communication contacts were directed toward him. It is interesting to note that when the manager was the receiver of information (or communication), the contacts made from outside the firm accounted for almost three times as many as those made from inside the firm. lll Tabular recording of communication flow. Table 14. COMMUNICATIONS TO: OUTSIDE FIRM: INSIDE FIRM: 5:90 w 2.30 1 1 9 5 ”83:93 90a 2 4 anew.» 00:33:— 3 1 2 2 >mauoflo~m8m 89.5 .H ‘ a mm .m Ema mama xmo .m *3. .on 5&3 sfiv .m 5:33. x3 .2 8.2800 m8€0 .335 «88830 g g « meUoEEm *5. .mm $8 .3. 4 g « « SEE mama 114 .mqoflmtwummno H30» moH mo mmmncmouoa m mm GBOnmuumummmamE 8&3 HID swam 88.2: MOM 30G :oflmufisflhfioO .vN 8.33m :8 .~ :8 .a 32195. *8 .v 98:5 3:53— 88830 a « SEE QOHDO *vm. H .835 .8955 .H. SEE mama Xmo .a a a: .m 223mg :3 .3 £850 .935 88890 x o mmmo 1383 N .02 nowmcde MOM 30:me mo ”.03 28350.73va 03836.90 .t UQOmh. {PVOAJ‘ Lu_ow sooo—oo‘.oassch H 10 0.00—I0000ecoop c;;.v .a ma unq>nwega .uzm:_avcu wu~»r:.>n ,(:<:rt.ox ooo.¢am coo.cov cca.ma- coc.ecn ou-.~Hp as..ac~ ass.oew s-.xmfi cea.«_- nooooooooowoooooooos— coo-coconoooooo‘eon-oooowoooqoo‘0.....—.ooooooo£~.ocoooouo~os.s oo.c . ac.oa . M H _ . _ .o.o~ . w o . L no.9» . m o o ac.ov . o o ac.oo . o o o ao.oo . Hnmw o o .o.o~ . o . qwaaufimn no.9: . nnmmmwxxxxe . dad-aHa-aaflam-x-«x . wfianwaaauaaawaafiaadd-afia.Hfinnnnxxxxxxxx. ax“ .«aaaaaaaaafiaaafiaafimmnmmmmmnnmnmmmmmmaxxxxxxxx«x4. . a-um oo.oo .xxxxxxxxxxxxxxxxuxxxx.xxxxxxxuxxaxxx . ms“ . x . xxra . eeexoecoeaeeaeeaolhh xxuxxwaa uc.coa 04414-404444444444- xxxauamnrma noooowoooo—ooooeoooo—moooooooo—oowoeoQooncoco-mow.—ooaowmooo—oooeoeou.pogo-ooooo~o.... soa.oum coo.co¢ cam.mav soo.non oo-.~4m as..«c~ ooo.onm ca~.xm« a:«..sa ea n .voa.o.a a“ n .wv;.a.; «m u .Ha=.g.au. fine» cuoo.o ..co«.o.o oncu.o nacno.o.: mn-.a nasao.c.=.. mJ-azu»z. COVVQBG n >33 3...»: (min... oooo.vaw u aor .3303 m .02 umwmcdfi yam madam? mo ”.on 20359.33“. 0>335850 .NQ vuswrm oaa.ono con.mnv ccc.oon goo.mfin oo~.a\a Jan.o~¢ .om.nn. u»...». .cv.~. ._...: cae.« ~ooooooooo—oooooouoo—oooo#oooo~oouoooooo—oo._onowo—oooonoo..—oofioooooo—ooon.ooo.—o .o.....—”....—....~ ao.c . c¢.o o o o .o... . co... . oc.o~ . oo.7n o o .o... . .o... o o no.9. u on.ov o o cc.on . co.om o o o no.0. . no.9: o o . ao.os . «an co.an . aaxxx . Hanna . ”ammmmnxxxx. . .o.oo . . aaaadadaflfifimnxxxwx . .cn.n¢ . Adannnmmxnmmmxxx .u. xxx . . udadaamm_xxxxxxxxxxxx ... rm. . .«H.....uuufidadua.........d.....mmnmmmmxxx.xx x . uo.oo .xxxxxuxxxxxxxxxxxxxxxxxxx‘xxxxxxxxxxxx xx. .ca.ao . xxmd . . xxx«aa . . .y..... xxx..xmmaa . . cc.ooa 0000......aoocoouoccouoac xxnmmmnmn .cn.oc« moooowoooo~ooooooooo~coo-cocoonooJoooooo—ooo.anon..—so.abut...—oo.oooooo—oOoooo.oo_on.o....omooo..o...o ooo.onv can.mov cco.oon aoo.man .ow.fl~m can.c~m pac..o. as"..nfi .nq.wm ‘p.... cu-.~ on - .~oo.a.a o" u .tva.:.; «w u .afin.:.su- .w»._.1 m< u4.>nm»r. “won...“ n~cou.o.c p.moaé ".9305; 1-4.o 13?...Lzuu w4u>zu73 .._ozw.:u...r.u 1:53;? ..:....rorx.;... «nan.ns u >ua awn hmnfi.wg u 7.3) q l u I oooc.nvs u mcz 3.30.3 w .02 umwmsde MOM madam? mo ”.03 £03.52»va 2.335980 ooo.nom coo.vam coo.mnv noo.o«¢ ooo.~nn can.mow ooo.onm c::.crfi rcc.um~ (::.ac g::.v uncooo.oo.o—ooooooooo—oo‘ovooouooJoooooo—4.ooouooo—oooonAcom—.oooooooo— cocooooo—oo.......~.o....J.... no.9 . - no.9“ . o oo.o~ . o oo.on . o o ao.ao » o . co.=a . o o o .o.oo . o . ao.os . . mxxx cc.oa . «Hnm.x . undaauxxxx uxm. . .........dvmmmmxxxxxx .m. . .. a......mnmmnmixxxxx...xxxxxx x... .. y. .o.oo .auuundwaaanuanaa«unmnmmnnxx xxxxxxxxxxxxx .... xxx. .xxxxx.xxxxxxxxxxxxxxxxxxx xxyma. . . .xmfia . cc ccoccccwhkbbbl xnm»,_ voud.° ..ooa.o.o cwxo.o n.9no.o_o mooo.o w.g«o.o.c.. aJ.>xu.z. muzmcfiuscu n.~»m:.>m .rzcccr;fiy mrcoqnc « .ua mm" nvro.\; u 7.ms oomc.oom - wap 3303 m .02 nowmcme new madam? mo .03 cofisniums 9,335930 . ooo.ooc con..vc aoc.mon oom.vcn o.o.omn o:m.uva can.o,fi 130.:nH =0:.«n. "momfi ”Waco”. coo-co—ofiooo:ooo— ~osooooooouooooo...o~ooooooooo~floooooooo—ooooooo4anoooo.ooo.-oooo ooo‘oooooooaupooo no.9 . p . co... . . . . oo.=m n . a 99..» n - - oc.cv . . a . oo.on . . A . Hafim . mmxx ao..o . a..x . .nmx ” .axxx awn ao.ou “ .mxxx . naaanx . aawnnn.‘ .H.an.xx. . co... . ......H¢..........Hmnrx4 xxx. . .a.............¢..H.nnnmmnmmmnmn «moxxxxxxxxxxx .mn.. .«annanunuauaa«Hamnmmmmmmmmnnmnmannxxxxxxxxxxxxxx xxx .xm .mannnmnnmnnnxxxxxxxxxxxxxxxxxxxaxx .mm. oo.oo .xxxxxxxxxxxx xx... . xxma . cut-ccooccoo Kmdu . ocoocccooccccccccoo pl: “xxx‘na 00.96" p41CIC1ClC€{1 xthh d o # uooooooooo~ooooooooo~assoc-ooou coo-oooo—ooooo.o omoaooooooomooooo.ooo—cooooo..o~ aao.ooq can..cc ooc.non can...» 9.3.0.. can...“ a...o;. zan.=n. a...m:. v. a .~.«.o.a o. u .~.g.e.o o. » .).a.o.u-- cm. «voa.o u.ooa.c.u snaa.= «.ono.o.: mmn~.a «.oao.c.g-u 14c>ru_z. ugzur~u7ru o_w mmcn.mm u >u4 awn meow oooa.moc u waz.z 5.3:.nn "maoum ,._cu: am. » ongooo—ooooooooow :am.mn uou.n h0g1 no ,Jr>4u_,~ pm.z.m >pr.auf_xv u.1. 9930.3 o .02 umwmame MOM mfimcmwm mo 909m 983959.3va 999935930 .mQ vuswwh 999.nnv 99m.9om 99¢.9vn 999.999 999.999 999.o9m 999.999 99¢.909 999.90 999.\v 999.9 ~o9ooooooo—ooooooooo—99~999999—99999cooomoo:oooooo~ooooooooc~99floooo9oa..ooooo.o_oo oo-ooomood...oo.~ 99.9 n .9.9 9 9 99.99 H 99.99 9 9 909°N H :09DN 9 9 99.99 H 99.99 9 9 99.9. n 99.9. 9 9 99.90 H 99.9m 9 . 9m 99.99 H 999 99.99 9 fix . 99» . «adamxx 9, 99.99 . «aanmxxx . 99.9. annxxxxx . . fuaaxxx .9 9x9 . 9 waxx co. xxgu . 99.99 . «99: .9 xmfi .99.:9 . 9999999999999... 9x99 . . 9999999999999.mnnxxxxxxxxxxxx 9x999 . . 99999999999999999999999«Huamnannnnnxxxxxx9999x .«xnm99 . 99.99 .«annnnnnmmmmmmonmmmxxxx.xxxxxxxxxxxx .. x9xm49 .99.99 .xxxxxxxxxxxxxxxxxxx ...... xxxx9949 . . xLxmnaa . 9 «0.0... so. xxxmAdfi . 99.99. .--uu¢-4-«‘99<4<«... xxxxxxmm 9 9 .99.999 nooooocooomoo-oooooouoo90’Cooonooooooooomo. cocooowoo-ooooo.~oo~cooomodooooooooowno oooo..~o.o..ooou— 999.99. 999.999 99....» 999.999 999.999 999.999 999.9.. 999.999 999.99 99.... 999.9 99 . .999.9.9 9N a .999.9.9 99 . .999.9.9-- 99..99a 9. 949.99.:_ 9999.9 9.999.999 9999.9 9.999.999 9999.9 «9999.999.. 99.>xu_z_ 992999.299 99m.m::.m >99:3.X99. homo.nn u )uo :_n 9990.99 u zqu; oooo.®Nv u war}: 3990.0. "muoun Zamgu: Pym. n mdrfh ._ rmhiyi. 153 9 39289...-99 9.99.99 . 9 9 9 ~9 9 .oZ ummmnde 90w9m9d99m9m mo 99on 9909995999399 9939399999990 999 999 99. .99 999.999 9 .999 999.999 999.9. 999.99 999.99 .9... ~9;999999.—9 99.9 . o9.9fioo—.99999999~999999999—9fi9999999_9.99.9.999.9.999‘999—9999999.9_9.9do9.9.— 99.9 . 9 9 Hr . x 9 A»... 99.99 . nx 9 9 9 99.9m . 9 9 9 99.99 H 9 9 99.9v . . 9 9 99.99 . 9 . 999x . 999x . 99999 99.90 . 9nn99 . 99999 . 99x99 . 99nx 99.9. . 99999999999x. . . awrmmmnnmxxxy a x . 999mmx9xxxxx» . x>xm . aunxxxxy xnm9 99.99 ” 9999999999939;9 .9999 9999999999999mnmnnnnnxxxx9 .99. . 9999999999999999mmmmmmmmmnmxxx9xxxxxx o .9. 4(499 .999mxxxxxxxxxxxxxxxxxxxxxxx9 .999 99.99 .99.. .9999 . xxxxxxxxxxmn9 9 0000.0... xermnnnmmanH . .. xxxxm99999999 99.999 .9 xxmmm99 ~9¥9999o9ohoo-ooooco—990999999—999999999~9.cocoa-o_-rooooco9—ooo999999—unocooouo~oooououooh9.9.9.99.— 9999mo~ 990.mow 990.09N 99¢.99N 99m.mn9 999.0vw 990.099 990.9) 999.90 .9m.~w 999.9 99 9 9999.999 99 u 9999.999 99 u .999.9.9-- 9n..991 9. 994.99.:9 9999.9 u9999.999 v999.9 "9999.999 9999.9 "9999.99n.- 9999z9_z9 999uc9...u u.x»m::>w 9.7.3999.v 9999.99 9 999 :99 9999.9. u .49. 9999.999 n 9924: 9999.99 umcoum 9999». 999 u m99.99 ._ 9.9;.; ca ondmwh 99.99 99.99 99.99 99.9. 99.99 99.90 99.99 99.99 99.999 154 49990399009: .m> 9.3.90.3 w .02 umwdcme no.9 madam? no 90999 9903599999399 2999339950 SQ uudwwm 999.999 99m.m9n —.9...9...~.9A 009° 9 9 9 9 99.99 H 9 9 99.9m “ 9 9 99.99 . 9 9 9 99.99 . 9 9 9 99.99 H 9 9 99.99 H 9 9 99.99 H 9 9 99.99 n . 9999 .999999999mmmmn 99.99 .xxxxxxxxxxxxxx . . 9 «can D 9°.oowlllluuuuqaauc —999.99999_.99 999.999 999.999 99 o~99.9 u9999.c.9 9999.999 999.999 999.999 9999.9—9999999.9—...9; 99999999999999999999n9 mammmmmxxxxxxxxxxxxu: 390 «99»; 99:.9. o rm(n. Ophh0_l Dd o_a9wvt9m .ffloflpv D ton-fic \ _u><9 .9 l" ~\J 99;. m =......-..— coo..o~ooooo9ooo—999999999—ooooooooo~oo-coooo9.ooooooomoquoo9ooo.o—oooooocoo..999oo.9o— 99n.\99 on..rm 99:.m 94.91u999 bracacr469 u ,un :92... 99.9 99.99 99.9¢ 99.9m 99.99 99.99 .99.99 .99.99 .99.999 1555 .Sdoflmuomfi .m> H3503 o .02 Hommcwflu new madam?“ mo ”.on nofldflhmg 9,335830 cco.nov occ.mon coo.m~n 99°.mnm coo.n.~ po°.m°n occ.mon w oao.n~“ ooc.rm moc.mv aoc.m uuoooonoo.—coonooo-ouooaooco.cage-oooooo~0ooco_vcoo—o“..-coco—.oo‘ooooo—ooooooo.o—woioooooowodo.onooo~ oe.o . - - - . o n . ao.on . o o o oc.o~ ” o o 90.0» . . o . . a oo o. o a . oo.om ” o. . oo.oo . c a . «xx oo.ou . «max . unamxxx . adaxxxx . "ddfifln xxx . .. ao.om . “adamxxxxxx .. . awudadumnmmxx C. . adudaumemmnmmmmxxxxxx .. .“auflaufiaauafl"umuuufladawfifiduawanmnmnxxxxxxxxxxx"xxx ... xxxma oo.oo .nxxxxxxxxxxxuxxxxxxxxxxxxxxxxxxxxxx ... xxnma .x xxxndad . COCCOCOC x XXded . cw... xxxxxxm nu 90.90“ .oqdIdddddddddddddddddddddddl xm uuwd uoo...oooo~choc-coconooocoooo.~ococooooo—ooooooooonumcoonous—o.coo-ooo—uoooooo..woocoonowcuoooo.....~ coo.nov coo.mon o°=.n~n occ.no~ ooc.mv« cae.nom oao.mow poo.m~d ooe.mo mac.mv coo.n “a . .«««.o.a o“ . .nmc.c.o nu . .ouo.a.o-. ou»»o;a m. m4.>am»z. ovoo.o ..co~.o.u ncoo.o ..ono.o.o "nu“.e ..o«o.=.o.. m4.>¢m»z_ auymn_uzcu u_¢»m1x>m >c¢cooxgoy «ana.us u >mo ohm mmvo.ns u z 3303 OH .02 ummmcdfi new madam? mo uoa moan—59333 0333550 .oQ 0.3th coo.o~o oom.noo coo.hnm oom.o~v ooa..:. oom.snn ooo.ann oofl.co~ ooc.¢nd can.dn coo.n ~..coo....—.H.H.....ncoooogo.on..ooooooan .ooo...o~o.ou_roooooo—oooovco.o—.ooooo_ioo—ooouo.oHA ~o....oo...~ oo.° H x oa.o o o ac.c« ” co.c~ . . oc.c~ H co.o~ . o oo.on ” oo.on . . oa.ov H oe.oc . o 99.0“ n oc.om o . oo.ao ” ce.oo o . ao.oh U c=.on o - oo.om H «flag c°.oo . «dfiflfiflaflanxxxx «dummmmn=xxxxxxx . Hflfififinfifldfifi"Hfa“a“fifififlafladaaaaadafiauflfldfi‘"aaaafifiafiamAammxxxxxxxx oo.oo .mxxxxxxxxxxx.x’xxxxxxxxxxxxxxxxxxxxxxxxxx.x.xxxxxxxxx.xxx ..w xxxm .oo.°o .x . xnfin . . xv." o . oooccccwu firm“ . 99.00“ .caccccooccccococoycoccc-ocococoaoocac-cuu 0 xxx: « .cc.oou H..-oooocuuo.a...oo-uo-oooocoonooo...ocow.on...ooouoooooooo-ngoooooo.o—ooooooooo—oo-ooooooh-ooooooocu ooo.e~o com.nco occ.nnn oom.os. ooo..o. can..nn ooc.«s~ can..o~ oce.ona mom.¢n eac.m a" . .n««.o.o o“ . .onc.e.o nw . .adc.o.o.. ow»»oaa m. ma.>¢m»z~ nvoo.o u.c=fi.c.g ovoc.c ..cmc.o.o omda.o u.cac.c.o.. m4.>aw»z_ mo:mo.uzcu uum»wxx>m >c¢cgozgoy cowo.om - >ma chm «omo.os a 24m: coco.moo n wo24a, coco.om awaoum zcucm: mam u m;ax<¢ 2. «main; 157 .Smofiouoofi .m> H9303 : .oZ ummmcmg new 39%? mo ”.03 ”.8399“:va 03338.90 ooc.n~. ooojcwv cow. an oom.nnm oov.~m~ Hococoooov~onooooooo~oooog...o”.-00.0.00”... . g on o . o o no.9“ . . o . oo.o~ ” p . oo.on . . u . oo.av . u - - co.om. . o a . oo.oo . . . . oa.on ” . . oo.oo . . fidflflflfidfiflfidw .aaaaflafiauanuwarafiafiafiauauaaaaaaaimmmmmmmmmmxa .mxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx oo.oo .x o . \v . Ill couoo:ococacoc00c oo.ocn acocoocccCIAOOCccc —-oo.oo o—oooooooo.aocooonooo—o o....oo~ooov coo.nnv coo.o~. oo~.con oom.nnn oov.hmn ya . .noo.o.a ma snea.o u.ooa.o.u noafl.o -.omo.o.o unv«.o ocoo.vov u wo2am»z~ n.oao.c.a.. m44>mm»z_ uoymc_uzco ufimpmxz>m >cmrucx40y oflmn.vc u >ua chm wmcc.~md u zqmr umaocm z«_cwz and u m4asN8< .No< 60¢. .02 893% £26580 .NN NNNNN 160 63:080. 0v." “(ON—4 C-I 033:5 N ~20 N wZOHPUDumZH Ove/NU H.330 083. £023.. mamanNonNNm 5:5 093 R. 03330 N90 mN.m ENNONNN 202.3. 0598.0 .850 magnum soup. N wagon 00230 3.232096 9‘.“ N $5.30”?— < 5% 02 DEED nonNom 3 000.5 332 mm. Eamomm < 0528mm 033:0 153m 03%? 153m Uzuzuéh. JU< .02 130.? z .6< ON N. o .0 Z NquNNNNE m m N .— >um>mu0< .Nowwawfi Non 3.33"; Nun 30395030 «ENHEwNa no 33832 .mm .3an 161 3 8H 3N c N NN MN 9. N mm HN NH 8 3489 9348 3 .2 Hm N N HN H N HHHE HHH x33 H352 mm .3 8H NH Ham H H.N N e NH. H805“. .825 N HN «m H 9 H. o Hm 55.0 N N H N m N 8 H32 H 8&8 .NN NH H H N N H N m N on mHzHEmem H N N N H v 88TH 8on HHEW OH. H N N H H ”235 60 mm .o 8 H m N N m N H m EH29 magma H H nomuuuomn— .8 .353» on. N H H muEmzonHmmmmou 2. . N N msgmmha moH H N N m H H 73» £83 8&0. 850 H N 38 $336 $383 8&0 N H H 0390 firs SHuHHaaou N «flaw 5m? gaudugOU 3 .N NN H N H. m H. H 553 mzzbom x 62 H38. HH 2 N N e m H. N N H 5333‘ .U< 60¢. .02 Humwuaz £3388 .NN ~33. 162 3032890 08330 mo 8 SH .2909 mm .N Na .N E .2 HN .e mm .3 Na .NH 8 .NH Nm .mH. mm .0 mm .3 mm .8 5.5.0 mN .3 mm .m Ho .2 NH. .NH Nm .NH Na .2 mH. .N No .o. Nm .m «H .m NH .m mszHfimm 0N .v NH .3 No .N NN .N Hm .N Hm .mH Hm .m No .m 3 .N Hm .N H..... .2 .HMSSHH Nm .m 8 .H. 8 .H 8 .H. mm .H 2 .N Na .H Hm .N H: .N muzwazonHmmmmoo No .N «N .m x. .H NN .m H; .o wsmSmmE N2 N0 .8 8 .NN mm .mm mm .HN NH .NN 8 .mm 8.3 mN .3 mm .NN mm .3. mo .NN V203 2.50m NN .HVH NH. .m Ho .2 mN .3 an .H. NN .m om .N 8 .o NN .NH H.H .N 2 .m mzofiosmbmzH 0258 NH. .2 Nm .2 mm .N 8 .3 Ha .N 8 .H. NN .H. mN .o Xw .HH mo .2 mm .2 65¢ 0250mm mo .2 mN .HVH mod 9. .m NN .2 NM .H. 8 .N mm .N NH. .3 HN .NH 8 .HH 5.30% < .503. RE UEHEU 8 .m 2. .o N». .m Ha .o 3 .m 8 .m 8 .o 8 .H 5.50% < 0658mm NN .o X .o oszHSE. HH 2 m N N o m H. N N H 5H>H8< .oz .HonHHuE .umwmamE .me 33303 Hum maoflutomno 05353 no unmoumm .mm 223. 163 63:: 8a 33 * 80392030 «Pagan— mo 3 cog 290B 0 .Nm N. ..—m v .8 0 .0m m .mm w .No 5 Km N .ov a .mb m .mw as u .Om m .mV 0 .NN — .mN m .MN 0 Km 0 .0m w .mm o KH 0 .Nv Hakim m .mw m .m: m .3 0 .MN N .m a KN « .0m N Na 0 .m o .wa Egg. N .mH 0 .mm o .0” 6 .ON 0 .v 0 .on 5 .0m m .04.. mUZmQZOmwmmmOU m .NN m .2 adv o.o~ mag/EH7: no. m .VN m .mm m .ON w .ON o .3 u Km v .8 o .mm m .m m .«m 2&0; maSOm m .3. m .8 m .om n. .mu 0 .0 m .N o .w m .Nw N .m m .oa mZOHHUDMPw—AH 0E0 w .ma 0 Km w .o.«. v .mN N K.— m Km m g: m .Nm m 6H w .fiN ZOmHU< UZHQUMQ N. .mm m .mm a Km 0 .mH c .m v .3. m .VN 0 .mm m .0." w .ON qumomm < 584 BE DEFEND Dim adv m .fiN v .v m.w¢ OJ o.mN 5.58m < 020mm 0 .ON 0.0N UZHZSH. : o.“ m m h *w m v m N a >3>30< .oZ «cw—25.2 .Eomuaimmpo ofimiwvluuowmcug 3m figflom “mm #83 A8usfifi an 2:3 «539:6 mma~o>< .vm £93. APPENDIX F Table of Student's t Test Values 492 Enuufiomawu >13: n .1. 42,2 Xmlasofiawa Sagan u ... .333“:me «on u .mdu rm XHQZHanfioa 986 .~ “1.88 .N 83 .H 3mm .v *3me .N mNmm A 8mm .5 mmmho .u .a a a 38 .H .u. .n .m d $8 .0 .m d .m d .m .a mgz 92.4. ggfimm .m .a .a .a mmmu .o #380 .N omma 4 03.0. .m nuvo .N *oomm .H One .N géh a .a a a 33 .0 38 .N *Nwwm .H Tm: .N *3me .N 8mm J vmmm .m mUZmQAOmmmmmOU m .a a a N58 .0 .m .n .u d nmnm .o .m d .m d .m .n ggmmhzu mom .m .a a a ammo 2" .m .a .m .n Numm 2— .u .n .m d .m .n 595 mZHBDOm a d u a $35 .u .m d .a d 9.3 .o inmmmw .N omwm 4 am: .m mZOCUDmeZH 0520 a .a .m .n 33 .0 00mm .N *88 A whom .N gamo .N umwa .H nmum .m 203.94. OZHQUmD a .a m d #80 .o .m d .m d mam .o ".0."on .N «Rm .H Nomv .m Eamomm < PDOm< 299385 059.50 a .a .m d 3mm .n .m .a .m 4. 3mm .0 ammo .N *ommm A «mum .N 3.80% < 020mm a d m d 9.3 .o .m d .u d vmom .O .m d .u .n g .m d UZHZHEH 8 up mo .u Anew do .u we .u Ago» 3 .u ma .u Juan 53301.. a: 5: use mafiaeom 6&2 Wu :4. am.» m6 ammm 51.32 :3 9:5 3E SE .9 9.0 .35 .9 8.0 .35 .832, 86» u mmfiuvfim mo Snap. .E 03.3. 164 APPENDIX G Tabulation of Data from Prototype Instrument Serial No. 81 Table G1. APPENDIX G Tabulation of data from prototype instrument Serial No. 81. T Mean Interval = INI— (Total Elapsed Time) _ 2114 minutes ‘ 89 = 23. 75 or 24 minutes 1 sigmal/24 min. Mean Rate of Occurrence equal to Time (AT) Total Time (AT) Total Time (AT) Total Interval Blamed Interval Elapsed Interval Elapsed Signal Between Time Signal Between Time Signal Between Time No. Signals (T) No. Signals (1‘) No. Signals (T) 1 16 16 31 20 768.6 61 5 1331 2 49 65 32 28 796. 6 62 7. 5 1338. 5 3 13 78 33 56 852. 6 63 33 1371. 5 4 61 139 34 18 870. 6 64 2. 5 1374 5 20 159 35 2. 4 873. 65 21 1395 6 10 169 36 2. 4 875. 4 66 15 1410 7 82 251 37 7. 6 883. 67 15 1425 8 16 267 38 10 893. 68 23 1448 9 2. 4 269. 4 39 31 924. 69 46 1494 10 7. 5 276. 9 40 8 932. 70 21 1515 11 78 354. 9 41 33 965. 71 41 1556 12 2. 4 357. 3 42 33 998. 72 5 1561 13 25 382. 3 43 11 1009. 73 64 1625 14 13 395. 3 44 10 1019. . 74 44 1669 15 5. 4 400. 7 45 2. 5 1021. 5 75 110 1779 16 7. 8 408. 5 46 2. 5 1024. 76 2. 5 1781. 5 17 51 459. 5 47 2. 5 1026. 5 77 2. 5 1784 18 5. 4 464. 9 48 10 1036. 5 78 5 1789 19 13 477. 9 49 13 1049. 5 79 62 1851 20 70 547. 9 50 10 1059. S 80 81 1932 21 7. 5 555. 4 51 5 1064. 5 81 5 1937 22 64 619. 4 52 20 1084. 5 82 20 1957 23 5. 4 624. 8 53 36 1120. 5 83 31 1988 24 38 662. 8 54 5 1125. 5 84 13 2001 25 2. 4 665. 2 55 74 1199. 5 85 5 2006 26 15 680. 2 56 2. 5 1202 86 13 2019 27 2. 4 682. 6 57 15 1217 87 31 2050 2 23 705. 6 58 39 1256 88 15 2065 29 23 728. 6 59 16 1272 89 49 2114 30 20 748. 6 60 S4 1326 165