{In LIBRARY Michigan St. ,. to University This is to certify that the thesis entitled CrFHEAeTc-Eelsms 0F Percema) Elm/new , name. UNGEZTAINTY 73514 INTEZDEPaJDamE A-pO 774612. #94; K” I4! 5 am? Raw presented by flow: Amy Dan. has been accepted towards fulfillment of the requirements for Mdegree in MA) . I [M Major professor DateM/If ‘ 0-7639 1:: st‘ ‘/Q\§-rv; ‘ amomc av ' -. . lscmmnm we. ., RAQ‘.’ BINDE RS '. , L m M“ "amen". ”all“: 1.] I m . ighb‘ll ul..4“flll..‘-l.l.u.ll-r ABSTRACT CHARACTERISTICS OF PERCEIVED ENVIRONIENTAL UNCERTAINTY, TASK INTERDEPENDENCE AND THEIR RELATIONSHIP TO SUBUNIT POWER IN COMPLEX ORGANIZATIONS By Thomas Harry Dulz This study has two main focuses. The first is on the conceptuali- zation and identification of organization environments and dimensions of the environment. The second is on the relationships between task and environmental variables and subunit power. The sample consists of three subunits of sales, manufacturing and engineering in each of twelve manufacturing firms operating in similar macro environments and having similar technologies. A twenty-item ques- tionnaire was administered to the chief executive officer and the re- spective heads of sales, manufacturing and engineering in each firm for a total of forty-eight respondents. As the units of analysis are organ- izations and subunits, responses are pooled to reflect the shared per- ceptions of the executive group in each organization. Environmental components are specified in terms of systemic inputs. Based on a conceptualization of an organization as an open system, six systemic inputs are identified which are common to all classes of organi— zations. These six inputs are the avenues or "linkages" between the or- ganization and its environment and are: l) labor supply; 2) material supply; 3) capital supply; 4) product technology; 5) process technology; and 6) I: ponents An literat‘ been well work in t ceptions dynamic a: ‘mcertain‘ Ship is be The d on ”New firms in t3 cepthns 0 more Prefi is 8350013 There the Variom 2 Thomas Harry Dulz and 6) markets. The environment is conceptualized in terms of six com- ponents corresponding to each of these six linkages. An environmental dimension which has gained much attention in the literature is that of uncertainty. However, to this point, it has not been well identified nor operationalized. This study builds on previous work in this area and empirically demonstrates the relationship of per- ceptions of change and complexity to perceived uncertainty. The more dynamic and complex the environment is seen, the greater the degree of uncertainty associated with that environment. The strongest relation- ship is between the dimension of change and uncertainty. If an environ- ment is seen as dynamic, it is also seen as complex. The dimensions of performance and immediacy have a mediating effect on perceptions of uncertainty. There is a strong relationship among the firms in this sample between past organizational profitability and per- ceptions of uncertainty in the market component of the environment. The more profitable a firm has been in the past, the less uncertainty there is associated with the market component. There is a significant difference in perceptions of uncertainty in the various environmental components. Labor, material and markets are seen as the most uncertain components of the environment followed by process technology, product technology and then capital which is seen as the least uncertain. The environmental components which have the great- est sense of immediacy for the organization in teams of effects - labor, material and markets - are seen as highly uncertain regardless of the degree of change or complexity associated with them. These results in- dicate that perceptions of uncertainty are a result of the interaction of change, complexity and immediacy, tempered by past performance. I Peri roamen ta that, to _ | ennronm‘ others '0‘, having m remnants, complex a controlh draw pOwe identifies sequently Pendencie from the task dep e No 8 “certs 1n 3 Thomas Harry Dulz Perceived subunit power is highly correlated with perceived envi- ronmental complexity and perceived task interdependence, which suggests that, to a large degree, power arises out of the dependencies created by environmental and task demands. The more a subunit is seen as affecting others by virtue of its task activities, the more power it is seen as having in systemic decision areas. In relatively placid, stable envi- ronments, subunits whose environmental domains are identified as more complex and turbulent are in a position to be more easily identified as controlling organization dependencies rooted in the environment and can draw power from this position. However, when the total environment is identified as complex and turbulent, no one domain stands out, and con- sequently, no one subunit can lay claim to control of environmental de- pendencies. In this situation, no single unit is clearly differentiated from the rest in terms of dependency control and so, the focus turns to task dependencies as a base for subunit power. No significant relationships are found between perceptions of task uncertainty, task difficulty, environmental uncertainty, or an imbalance in task interdependence and the variable of subunit power. However, the subunit seen as most critical is also seen as the most powerful. More reported differences are found when organizations are come pared on the basis of environment type than on the basis of performance. Organizations experiencing relatively high levels of environmental un- certainty and turbulence also report higher degrees of task interdepend- ence, irrespective of economic performance levels. Consistent with other reports is the finding that there is more self-reported power in more profitable firms. CHARACTERISTICS OF PERCEIVED ENVIRONMENTAL UNCERTAINTY, TASK INTERDEPENDENCE AND THEIR RELATIONSHIP TO SUBUNIT POWER IN COMPLEX ORGANIZATIONS By Thomas Harry Dulz A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Management 1976 I. II. III. TABLE OF CONTENTS INTRODUCTION . . . . . . . . THE ENVIRONMENT . . . . . . . Conceptualization of the Environment The Elements of a Methodology Energic Input Throughput Energic Outputs Identification of the Environment Dimensions of the Environment Complexity and Change Uncertainty Environmental Characteristics and Uncertainty Summary INTRA-ORGANIZATIONAL POWER . . . . Power Influence in Decision.laking Autonomy Interdependence Substitutability The Environment Performance The Critical Subunit Summary HYPOTflflEES AND VARIABLES IN THE STUDY Introduction The Environment Hypothesis One Hypothesis Two Hypothesis Three Hypothesis Four Power dd-fi—D—bd \qum-Fomm'dm \N 23 27 34 35 36 39 42 43 44 44 46 46 46 47 47 47 49 an Table 0 Table of Contents (cont'd.) IV. V. VI. Hypothesis Ten Hypothesis Eleven Hypothesis Twelve The Environment Simple-complex Environmental Dimension Static-dynamic Environmental Dimension Perceived Environmental Uncertainty Subunit Power Task Interdependence Asymmetry Task Difficulty Task Understanding Criticalness Performance Summary METHODOLOGY AND DEVELOPMENT OF THE INSTRUMENT . Sample Organizational Subunits Data Collection Methods DeveIOpment of the Instrument The Instrument Summary RESULTS OF THE STUDY . . . . . . . . . . Discussion of the Results of the Study Performance Performance as an Independent Variable Variance in Perceived Environmental Uncertainty Power Index of Subunit Power Analysis Correlational Analysis Intra-Organization Analysis Criticalness Relationship With Other Studies SUMMARY AND CONCLUSIONS . . . . . . . . . APPENDIX D O O O O O O O O O O O O 0 Introductory Letter Response Cards Questionnaire iii 70 7O 73 73 77 78 80 81 84 91 93 96 99 101 101 106 111 112 114 119 119 120 121 E 10, ‘1? \m “nvi to b‘ Type Varia Chara« Punct; Summa: Part1, EDVlr‘t Uncer‘ Organ: AnalYE EDVirc Analyt b“I‘ll“: Analys Envir: Mean ( Envlr< COPra‘ comPOz C0rre: Organ: COrn ~ 01‘8811: IEQQ I Organ: l. 2. 3. 4. 5. 6. 7. 8. 9. 10. ll. 12. 13. 14. 15. 16. LIST OF TABLES Environmental State Dimensions and Predicted Perceived Uncertainty to be Experienced by Individuals in Systems Operating in These Types of Environments Variables of the Study Characteristics of Firms in the Study Emotional Responsibilities of Subunits Sumary of Results Partial Correlations of Static-Dynamic and Simple-Complex Environmental State Dimensions with Perceived Environmental Uncertainty Organizational Performance Analysis of Variance: Organizational Performance upon Perceived Environmental Uncertainty of Organizations Analysis of Variance: Organizational Performance upon Perceived Environmental Uncertainty of Environmental Components Analysis of Variance: Environmental Components upon Perceived Environmental Uncertainty Mean Organization Scores of Environmental Dimensions, by Environmental Components Correlations of Environmental State Dimensions by Environmental Component Type Correlation Matrix of Organization Scores Correlation latrix of Organization Scores Controlling for Organization Performance Correlation latrix of Organization Scores Controlling for Organization Environment lean Differences Between Organization Scores by Performance and Organization Environment Type iv List of l7. l8. 19. 20. 21. 22. 23. 24. l Cor Dis Cor Int, Tots Rank for Bank for . Mean Sales List of Tables (cont'd.) 17. 18. 19. 20. 21. 22. 23. 24. Correlation Matrix of Subunit Power Index Items Distribution of Subunit Power Index Scores Correlation Matrix of Subunit Scores Intra-Organization Ranking of Variables Total Ranking of Variables For All Subunits First-Ranked on Power Ranking of Variables for Subunits First-Ranked on Power Controlling for Performance Ranking of Variables for Subunits First-Ranked on Power Controlling for Environmental Type Mean Differences in Coefficient of Relative Variation Values for Environmental State Dimensions of'Manufacturing, Engineering and Sales Subunits, by Organization Environment Type INTRODUCTION The original intent of this dissertation was to investigate the bases of subunit power in complex organizations. A review of the liter- ature generated the following conclusions. 1) Power is a very complex phenomenon and there is little agreement on its characteristics. 2) In most research, power is treated at the individual and interpersonal level as an independent variable. Only two studies are found where sub- unit power is the dependent variable. 3) The "newer tradition" in or- ganization theory places a heavy emphasis on technologies and environ- ments. There has been little agreement on the conceptualization of an organization's environment or the elements comprising it. 4) Theorists suggest that power is related to coping with uncertainties stemming from technologies and environments. These conclusions governed the design of this study. The first task was to design a model of organization environment, which would not be organization-specific and could be used across other classes of or- ganizations in future research. The second task was to operationalize the concept of environmental uncertainty. Chapter One deals with these problems. Chapter Two deals with the issue of power, and specifically subunit power. Various dimensions of power are discussed and operation- alized, drawing on the most recent work in this area, and relationships to technological and environmental variables are suggested. The sample chosen for this study consists of twelve manufacturing 1 firms se as 3111111 10!! comp Because t ables yet would int: more diffj prised of Chapt 510:1 in C1: that all 0 Chosen bec 3hapter Th ”Plains h Chap: methOdOlos We: and 3139 Sauna :Poader or firms selected according to the following criteria. 1) They were to be as similar as possible in structure, technology and environments to al- low comparisons between them. 2) They were to be simple in nature. Because the study is exploratory, with the relationships between vari- ables yet to be established, it was felt that large scale organizations would introduce an element of complexity which would make the analysis more difficult. Consequently, the sample consists of small systems comp prised of three main subunits - manufacturing, sales and engineering. Chapter Three presents twelve hypotheses derived from the discus- sion in Chapters One and Two. It was not expected from the outset that all of these hypotheses would be supported by the data. They were chosen because they represent directions suggested by the literature. Chapter Three also contains a list of variables used in the study and explains how each is Operationalized. Chapter Four discusses the development of the instrument and the methodology used. The results of the study are presented in Chapter Five, and the concluding chapter, Chapter Six, reviews this data analy- sis, summarizes the study as a whole, and places the findings in a broader organizational context. CHAPTER ONE THE ENVIRONMENT In recent years, organization theorists have been asserting that the environment is a critical factor in understanding and explaining much of what goes on inside organizations. One of the earliest propo- nents of this theme was Dill (1958) who traced differences between the two Norwegian firms he studied to differences in their respective envi- ronments. Dill's contribution was not simply an acknowledgement of en- vironmental effects, other researchers had reported organizational- environmental interactions (e.g., Selznik, 1949), but rather, he was one of the first to offer a theoretical conceptualization of the environ— ment. He differentiated between the "general" environment in which all firms operate, and the ”task" environment which is unique to each fimm. He identified the task environment as being composed of four sectors: customers; suppliers; competitors; and regulatory groups, each of which are characterized as being relatively homogeneous or heterogeneous. Interdependencies were the focus of an article by Emery and Trist (1965). They differentiated between "internal intendependencies" -- processes within the organization; "transactional interdependencies" -- exchanges between the organization and its environment, in either di- rection; and "causal texture" -- those interdependencies within the en» vironment itself. This conceptualization is similar in many respects to Parson's (1960) "levels" of organizational responsiblity and control. Parson ' a spend itch actional tinguish environml eat was ( "turbuler field its lent fiel Trist 8rg reCOgniZe ment and ‘ decreasiné ‘I'erreberr.j argues for tional Var deseri bed 4 Parson's "technical", "managerial", and "institutional" levels corre- spond roughly to Emery and Trist's "internal interdependencies", "trans- actional interdependencies", and "causal texture". Emery and Trist dis- tinguished four types of environments according to the degree to which environmental components exhibited "system connectedness". Their inter- est was centered on the most complex of the four, which they called "turbulent field". In this type of environment, changes "arise from the field itself" not Just from interactions of the components. A "turbup lent field” is characterized by ”autochthonous processes". Emery and Trist argued that the firm they studied, a vegetable canner, did not recognize the fact that it was operating in a "turbulent field" environ- ment and was unprepared for environmental changes which resulted in a decreasing market for its product. In an often-cited article, Terreberry (1968) building on the work of Emery and Trist and others, argues for the increasing importance of the environment as an organiza- tional variable. She argued that a "turbulent field" environment best described the situation for contemporary organizations and that environ- ments were becoming increasingly ”turbulent”. She further maintained that the environment was the gg§t_important factor in explaining organi- zational behavior and hypothesised that "organizational change is largely externally induced". Katz and Kahn (1966) in their discussion of organizations as "open systems" echoed this theme of the importance of the environment, arguing that a necessary sustem component is the "adaptive function" which has as its aim."environmental constancy". Survival of the system, in their view, is dependent upon the success of the adaptive function. J. D. Thompson (1967), in his book Organizations In Action, which he call on the c nents. powerful takes th by envirc the found existence Whil. theoretic: anvil-owe: in his crz‘ 01' ”earn; emF’il‘ical Preceas by i“emulsion tion and ( ultra-org Parity in vironment their Btuc' their 8831; he calls a "conceptual inventory", bases his analysis of organizations on the concept of uncertainty stemming from technologies and environ- ments. He sees environmental influences as ultimately being the most powerful in shaping organizational characteristics. Stinchcombe (1965) takes the position that, not only are existing organizations influenced by environmental forces, but that environmental conditions dictate both the founding of organizations and the forms they will take during their existence. While the importance of the environment has been well publicized by theoretical writers, a review of the empirical literature discloses that environmental influences are virtually ignored in most studies. Hirsch, in his critique of industrial sociology says: "In short, while we speak of organizations as interacting with their environment (in theory), most empirical studies, by virtue of their design, continue to ignore the process by which this interaction occurs". (1976:5) lot only is the inclusion of the environment as a variable important in interborganiza- tion and organizational analysis, but it is also an important factor in intra-organizational analysis. It is possible that much of the dis- parity in research findings could be explained if the nature of the en- vironment were identified. For example, Hinings, et a1., (1974) in their study of the bases of organizational subunit power, found, in their sample, that production was the most powerful subunit in contra- diction to Perrow (1970b) who found sales to be the most powerful sub- unit in the firms he studied. Neither researcher identified the nature of the environment of their sample firms. It is possible that much of the variation in these findings could stem from environmental differences. In fact, the value of any study done in an organizational setting would be Gouldner ing the 2 increase: while Gue later com Organizat be avoide One face, is differenc son agree. At this s are "di ff 310x18 0f not been ' i-flg disc,“ (:OllceptuaJ identifice would be enhanced if there were some identification of the environment. Gouldner (1954) and Guest (1962) both examined the consequences follow- ing the succession of a new'manager in an organization. Gouldner found increased stress and tension fellowing the arrival of the new'manager while Guest found the Opposite to be the case. Gouldner (1962) in a later comment pointed out that the environmental circumstances of the organizations were very different, hence, any direct comparisons should be avoided. One of the difficulties researchers, such as Guest and Gouldner, face, is that although they may recognize that there are environmental differences which may have a bearing on their findings, there is no comp mon agreement or precedent on which to make environmental comparisons. At this stage, all we know is that circumstances surrounding each study are "different". We are not yet able to agree on the relevant dimen- sions of these differences. The identification of the environment has not been clearly specified in the literature. Accordingly, the follow- ing discussion will delineate some of the major issues of environmental conceptualization and suggest a methodology for environmental identification. Conceptualization of the Environment The generalized term "environment" encompasses an infinite number of elements, some of which, but not all, will be relevant for any given organizational analysis. The first issue to be considered is the ques- tion of.!ggt is the environment. A separate, but related issue is, once the environment has been identified, what is the nature of it. On the surface, this seems like a very elemental distinction, too trivial to discuss. tion has. J. D. ThAI . | ennronn: on some 1 recent ce typOlOgy mane a p; rists dic infitead 1 is an is: discuss. However, a review of the literature shows that this distinc- tion has not always been observed. For example, Emery and Trist (1965), J. D. Thompson (1967), and Terreberry (1968) all discuss the nature of environments, (in terms of "turbulent field", homogeneity, etc.) based on some a priori identification which they do not make clear. A more recent case is the article by Jurkovich (1974) who offers a "core typology consisting of 64 types", all of which are descriptive and as- sume a prior identification. While it may be argued that these theo- rists did not intend to deal with the issue of identification, choosing instead to introduce criteria for evaluation and describing, still, this is an issue that must be dealt with if a theory of organizational— environment relations is to be developed. The Elements of a Methodology The first question to be dealt with in environmental identification is the method one would use in identification. The following discussion proposes a method of identification based on the notion of an organiza- tion as an open system. It begins with the question: "If an organiza- tion is an open system, where, and in what ways is it open?" The ini- tial focus is on possible environment/organization links es, which then provide direction for searches out into the environment for relevant factors and also back within the organization for environmentally sensi- tive elements. It is assumed that the organization is an "open system" subject to the influence of environmental complexities and as such exhibits the following systemic characteristics: 1) energic input; 2) transforma- tion of energies within the system (throughput); and 3) energic output (Katz an course, three ju ganizati' discussm Energic J Org: ronment. the necee Sanizatic Volunteer is m 1388 and 8 (Katz and Kahn, 1966:19-20; J. D. Thompson, 1967:23—24). There are, of course, other characteristics which are descriptive of systems, but the three Just mentioned are the most relevant in terms of environment/or- ganization linkages. These three provide a guide for the following discussion. Energic Input Organizations import two basic categories of energy from the envi- ronment. The first of these is lEEEE - those individuals who possess the necessary skills and abilities to carry out the functions of the or- ganization but this category includes all forms of labor: consultants; volunteers; owners; slaves; etc. The second category of energic input is materials, both those needed for system maintenance, such as build- ings and supplies; but also those needed for transformation into the finished product, such as, ...a living being, human or otherwise, a symbol or an inanimate object. People are raw materials in people-changing or people-processing organiza- tions; symbols are materials in banks, advertising agencies and some research organizations... (Perrow, 1967:195). A distinction is made between humans whose contributions to the organi- zation is in the form of skills, such as teachers, social workers, or guards (labor); and humans whose contribution is their presence, such as students, clients, or prisoners (material). In order to insure system maintenance and survival, an organization must have an ability to procure these two categories of inputs. The means of procurement can be in the form of purchase, barter, seizure, or third party intervention. Control over the means of procurement can be seen as a form of stored or potential energy which affords an organi za bility i. I later ex ‘ speaking. Changed f ubiquity ronment t, 2) materi Orga conditIOn 1965:913) (1966:20) Throughpu nature or The I cessad is effect or the same t P organization a degree of flexibility. In a money economy this flexi- bility is enhanced by the possession of capital which can be stored and later exchanged for necessary labor and materials. While strictly speaking, capital is a secondary input, in that it is at some point ex- changed for other inputs, it is included in the model because of its ubiquity in this culture. An organization will be linked to its envi- ronment through the three categories of inputs just described: 1) labor; 2) material; and 3) capital. Throughput Organizations engage in a "pattern directed effort to alter the condition of basic materials in a predetermined manner" (Perrow, 1965:913). This activity corresponds to Katz and Kahn's "throughput" (1966:20), and J. D. Thompson's "technological activities" (1967:19). Throughput includes two elements: the nature of the material; and the nature of the process used to convert the material into an altered state. The manner in which the nature of the basic material to be pro- cessed is defined will have important organizational consequences. The effect of this is seen clearer when we compare organizations which share the same basic material input but define the initial state differently. For example, Perrow (1965) has compared mental hospitals which he classified into two types: "custodial"; and "therapeutic", based upon their respective initial definition of the basic material to be pro- cessed -- people, and found them to differ significantly along a number of dimensions. Perrow (1967) and Rushing (1968) both discuss, at some length, organizational consequences stemming from the manner in which the material to be processed is defined. This definition, or an altere identifit technique Physical: to J. D. There 13 Process t istics (e PFOduct t 80mg exam 10 product technology, is heavily influenced by belief systems, culture, generalized knowledge and other environmental factors. The nature of the process used to convert the basic material into an altered state also has environmental roots. This process will be identified as process technology which Perrow defines as the "complex of techniques employed to alter material (human or non-human, mental or physical) in an anticipated manner" (1965:915), which also corresponds to J. D. Thompson's conceptualization of "core technology" (1967:19). There is ample evidence in the literature to support the assumption that process technology has an important effect on organizational character- istics (e.g., Blauner, 1964; Woodward, 1965; Perrow, 1970a). As with product technology, process technology has its roots in the environment. Some examples of technological developments that have impacted on or- ganizations in recent years would be: computers and systems analysis; PERT; organization development; credit cards; profit centers; plastics; and so on. Not only are existing organizations influenced by environ- mentally rooted technologies, but as Stinchcombe (1965) has argued, the influence affects the founding of new organizations. In addition to the inputs of l) labor, 2) material, and 3) capital, an organization will also be linked to the environment through 4) product technology, and 5) process technology. Energic Outputs Outputs are the "products of the system". They can be intended re- sults of the process technology - automobiles rolling off the assembly line - or they can be unintended results - air pollution. They can be tangible, as in the case of salaries, or intangible, such as psychic satisfec of funds governme: For those whi through-p outputs, possible Flex rela "11311 the . (Emery an. enviromne: 'ith the : as the aui hots of E “blishmam te°m°l°si The e 11 satisfaction; voluntary or involuntary, as for example, the distribution of funds to stockholders as dividends or distribution of funds to the government in the form of taxes. For our purposes, the systemic outputs which have relevance will be those which, in some manner, affect either the energic input, the through-put, or both. A difficulty here is that given the wide range of outputs, those factors which will affect the organization are often im- possible to identify at any given time because of the tenuous and com- plex relationships between cause and effect. This is especially true when the outputs are mediated by other factors in the "causal texture" (Emery and Trist, 1965), or "institutional level" (Parsons, 1960) of the environment. For example, one of the "outputs" originally associated with the introduction of the automobile was the reduction in pollution as the auto replaced the horse. Much later, the "discovery" of the ef- fects of air pollution stemming from the auto has led to government es- tablishment of standards which have affected both the product and process technologies of the automobile companies. The effects of outputs on the organization will be felt through one or more of the five inputs previously discussed: labor; materials; capi- tal; product technology; and process technology. Hence, a model of en- vironment/organization linkages could stop with these five components. For organizations where one or more of the end products is converted by sale into money, then product disposition, or the market component of the environment becomes a generalized predictor of future effects on in- puts. Consequently, in those cases where the conversion of output into money can be identified, then the market component should be included in analysis. For many organizations, especially those in the "non-profit sector" users of Donna (1 must tai. inputs. An were pro tne sysze tion is Chanée in In t: .have beer) ThESE are: Emil; t W of “Wm ‘ open to En. 12 sector" of the economy there is no clear cut market as in many cases the users of the organization's outputs do not supply input resources. Downs (1967) does an excellent Job of showing how governmental agencies must tailor their outputs in terms of their efficacy in obtaining system inputs. An alternate conceptualization is that used by Hinings, et al., where product disposition is seen as a "demand" and hence an input into the system, rather than an output (1974:28). Whichever conceptualiza- tion is used, market as an input or output to the system, there is no change in the manner in which this concept would be operationalized. In the preceding discussion, six environment/organization linkages .have been identified based on an open systems model of organizations. These are: those having to do with inputs: labor; material; and ca ital; those having to do with throughputs: (product technology; and process technology; and finally markets which have to do with one form of output - product disposition. All classes of organizations will be open to environmental influences through one or more of these six. The advantage of this linkage model is that the effects of virtue ally any environmental factor can be traced through one or more of these six linkages and gives a more precise delineation of organization- environment interactions. As an example, consider one of the most ubi- quitous factors in a finm's environment today - the government. Equal employment and minimum wage legislation affects the labor component of organizations. In the sample of firms in this study, one of the often mentioned environmental factors was the OSHA regulations which affected the process technology of these firms. Another governmental interven- tion which affected the process technology of these companies was the introdut' ment of fected t cized cs changes resisting cess tee: 8180 eff; try, one Performan cess tech; ment is a: One ( Ferment a: 13 introduction of clean air standards and regulations of the state Depart- ment of Environmental Protection. Consumer safety standards have af- fected the product technology of’many companies, one of the most publi- cized cases is that of the automobile industry, resulting in product changes such as lowerbemission engines, restraining harnesses, impact- resisting bumpers and so on. In this instance, both product and pro- cess technology is affected. Government intervention in these areas also affects the market component, in the case of the automobile indus- try, one of the results has been the switch in advertising emphasis from performance to gasoline economy. When these changes in product and pro- cess technology require capital investments, the capital supply compo— nent is affected. One of the things which must be considered in organization-envi- ronment analysis is the fact that not all organizations respond in the same manner to changes in the environment. It is difficult a riori, to forecast how environmental changes will be identified, and conse- quently what strategies of accommodation will be pursued. For example, a government increase in the tax on inventories may be identified by one firm as requiring a change in process technology - the introduction of a manufacturing process which results in a lower inventory level, or the introduction of a management science technique of inventory control with the same result. Another firm might respond to the same change in the environment by a change in the marketing area - a change in the pro- duct mix, drcpping those products with high or costly inventories, or increasing the price on the products. The point is, changes in an organization's environment become rele- vant in terms of the responses made by the organization in question. Not all ronments interact_ will con Hot only of envirt union to or class Organi 23 t Enviroan Ident'f' % Tnere Eaniza ti 0:: 14 Not all organizations will respond in the same manner to the same envi- ronmental change. Consequently a model of organizetion-environmental interaction which facilitates the analysis of organizational responses will contribute to a better understanding of organizational behavior. Not only does this linkage model allow the identification of the effects of environmental factors, but also provides a frame of reference from which to determine the relevant environment for any given organization or class of organizations and allows a measure of comparability between organizations. This model will be used in the following discussion of environmental identification. Identification of the Environment There are a number of approaches one might take when studying or- ganizational environments, the choice of which must be determined by the aims of the researcher. Osborn and Hunt (1974) propose a typology of three categories of environments: macro, aggregation and task. The 35252 environment "is the general cultural context of a speci- fied geographical area and contains those forces recognized to have im- portant influences on organizational characteristics and outputs" (1974:231). The most noted work in this area is that of Stinchcombe's (1965), who examined macro environmental variables such as literacy, ur- banization, schooling, political characteristics and organizational den- sity and traced their relationship to the formation and maintenance of general organization types. The value of Stinchcombe's work lies in the demonstration of the effects of macro variables on organizations in general, and is suggestive of some directions empirical research might take. The ronment constitu This cat; level" (L the E Ihich is This para (1967), L The .? ables m1. eats. lh; t‘3I‘lls of g enviromzler cussed. g the macro engage in £180th tn identifier: 15 The second of Osborn and Hunt's categories is the aggregation envi- ronment which they define as "the associations, interest groups, and constituencies Operating within a given macro environment." (1974:231-2). This category is similar in many respects to Parson's "institutional level" (1960) and Emery and Trist's "causal texture" (1965). Finally, the‘tggg environment "is defined as that portion of the total setting which is relevant for goal setting and goal attainment". (1974:232). This parallels the definitions adopted by Dill (1958), J. D. Thompson (1967), Lawrence and Lorsch (1969), Aldrich (1972) and others. The preceding categories can be viewed as a pool of potential vari- ables which a researcher will choose to investigate based on his inter~ ests. Whichever variables are selected, their choice must be made in terms of a-fgggl organization and that organization's relationships with environmental elements through the six linkage points previously dis- cussed. Kimberly (1975), for example, reports a relationship between the macro variable of increasing "social responsibility" and the in- crease in income from.grants of those rehabilitation workshops which re- flected the changed societal values. This is an example of a study identifying an element in the "macro" environment by means of its ef- fects on one of the organizations inputs, in this case, that of capital. Emery and Trist's (1965) conceptualization of the "causal texture" was based on their note of changes in the market linkage of the firm they studied. The model of environment/organization linkages can not only be used for the identification of the "relevant" environment but can also be used in intra-organizational analysis such as that to be discussed later. In intra-organizational analysis, the focus is not on environmental 16 elements, but rather on the adaptations and responses the organization makes to the environmental elements. To facilitate this analysis and allow inter-organizational comparisons, what is required is a conceptu- alization of the environment which is applicable to all organizations being compared. Conceptualization of the environment in terms of the six components discussed above satisfied this requirement. For the pur- poses of this study, the term "environment" will be taken to mean those factors, outside the boundaries of the organizations studied, which are identified by the respondents as critically affecting the organization in the areas of labor, material, capital, product technology, process technology and markets. Dimensions of the Environment Organization analysis requires not only the components comprising the organization's environment be identified, but also the nature, or dimensions of that environment. The following discussion reviews the environmental literature and catalogs the themes common to this litera- ture. Complexity and Change lost descriptive treatments of the environment in the literature center around the dimensions of complexity and change. Change is the major theme of Terreberry's (1968) discussion of the increasing impor- tance of the environment as an organizationally relevant variable. Tosi, et al., focus on "range of fluctuations" (1973:30). Osborn and Hunt choose complexity as "an important, if not the most important vari- able in the environment" (1973:233). Emery and Trist (1965), Lawrence I and Lore (1974) al complex: her art "eta tic- respec ti 01’ simpl. 9w} 17 and Lorsch (1967), J. D. Thompson (1967), Duncan (1972), and Jurkovich (1974) are just a few of the theorists who have identified change and complexity as major dimensions of the environment. Duncan, in an ear— lier article, (1972) has suggested the labels "simple-complex" and "static-dynamic" to represent the dimensions of complexity and change respectively. Rather than create a new terminology, his identification of simple—complex and static-dynamic will be adopted for the purposes of this study. Uncertainty The concept of uncertainty is a major foundation underlying the work of the theorists in the ”newer tradition". Crozier (1964) was one of the earliest to use this concept in accounting for the power of main- tenance engineers in the French factory he studied. He equates uncer- tainty with lack of predictability (1964:109). J. D. Thompson suggests a situation of uncertainty exists when there are "more variables than we can comprehend at one time, or that some of the variables are subject to influences we cannot control or predict" (196736). Weick, in dealing with the same issue, substitutes the term “equivocality” which he identi- fies as the range of "possibilities or sets of outcomes that Eight oc- cur" (1969:40). Quantitative decision theorists differentiate between conditions of risk, where the probility of outcomes can be calculated from past events, and conditions of uncertainty where the probabilities must be estimated. While risk is a single category, uncertainty is a continuum ranging "from near accurate estimates based on objective experience to an extreme case in which no knowledge exists" (Archer, 1967:455). La three e1 fect re; Responds uncertai Lorsch's ations o Hinlngs, Variabil OI‘ Varis (1974:2& "uncerta temativ the fires herent u One identifi environm 18 Lawrence and Lorsch's conceptualization of uncertainty consists of three elements: "clarity of infbrmation: uncertainty of cause and ef- fect relationships; and the time span of definitive feedback" (1969:28). Respondent scores on each of these measures was combined to get a "total uncertainty score". Tosi, et al., attempting to validate Lawrence and Lorsch's instrument operationalized uncertainty as the "range of fluctus ations of revenues or expenditures" for the firms studied (1973:30). Hinings, et al., assumed that uncertainty is related to unpatterned variability previously experienced, defined as "the degree of constancy or variability in three elements, trend, range, and regularity" (1974:28). They base this conceptualization on their assumption that "uncertainty is a lack of information about future events, so that al- ternatives and their outcomes are unpredictable .... It is assumed that the greater the variability previously experienced, the greater the in» herent uncertainty" (1974:27). One of the more ambitious investigations was that of Duncan, who identified uncertainty as: "1) the lack of information regarding the environmental factors associated with a given decision-making situation; 2) not knowing the outcome of a specific decision in terms of how much the organization would lose if the decision were incorrect; and 3) ins- bility to assign probabilities with any degree of confidence with regard to how environmental factors are going to affect the success or failure of the decision unit in performing its function" (1972:318). This con- cept was measured by a lZ-item Likert-type scale from which Duncan con- structed a "total uncertainty score". The common theme running through all of the conceptualizations dis- cussed above is that of uncertainty as unpredictablenesa. The more uncertaf Ofthe responds bility f it seems uncertai study, 1 much 11 Du: dynamic Tdainty the dim unceI‘ta differe 19 uncertain the situation, the less one is able to predict consequences. or the twelve items in Duncan's uncertainty measure, six had to do with respondent's perceptions of predictability. The Spearman-Brown relia- bility for this scale was .91 (Duncan, 1971). Based on this evidence, it seems safe to assume a major component in the conceptualization of uncertainty is predictability, consequently, for the purposes of this study, the dimension of uncertainty will be defined as the degree to which it is felt an outcome can be successfully predicted. Environmental Characteristics and Uncertainty Duncan found that ”individuals in decision units experiencing dynamic-complex environments experience the greatest amount of uncer— tainty in decision making" (19722325). Duncan's basic argument is that the dimensions of simple-complex and static-dynamic are components of uncertainty. "By considering the interaction of these two dimensions, different states of the decision unit's environment can be identified. Once these are identified, predictions can then be made as to the degree of perceived environments" (1972:320). Duncan has constructed a typology of environments based on these assumptions which is reproduced in Table l. Duncan's data supports these assumptions. In his sample of twenty- two subunits, he found that those units which perceived a small number of components in their environment which were basically unchanging (Cell I) perceived the lowest amount of uncertainty in those subunits in his sample. Subunits identifying the largest number of components and perceiving the greatest amount of change (Cell IV) perceived the great- est amount of uncertainty. He further found that the greatest amount of apeoenoagcm Mo conga scene 5. e53 chemo maoamam as maguyfiuCH .3 encamwhaqem on 3 xunfieukeoab ve>«eokon~ manuauohm 326 ODOfimfimEfiQ Guam Hlufiofiflohgnfl A. M143 «DH 20 ownono Ho ooooouo Hennapnoo a na one anoanoaa>no no opnonoqaoo one oaoaoeh An someone one o» neasaao won one opnonooloo one ouovoem AN anolnoaa>no one nH opnonoosoo one eaoaoeu «o Hon-5n owueq AH apneepnooep oo>aoonom nose no aaoo omneno Ho oeoooun Hennaanoo na one anonaaaabno on» no ounonooseo one oaoaoom Am someone one on seaunao wens Ioaoo one eunonognoo one eaoaoem Aw unoanoaw>no on» no opnonon nice one oaopoeu no nonnnn Hanan AH ameaaaaoonp oopeooeom eon: aaopeuooo: "n Heoo oaeo one haaeoaoen nausea opnononaoo one esopoem An noneone ono on heafiaao won one oononoosoo one ouoaoem Am pnoanoaa>no one nH ovnononaoo one oaoaoeu no nonann owueq Aa Nmnaoauoonb oo>aooaom.son haoaeaoool «N Haoo mnawneno won one one oaeo on» aaaeoween nausea opnonooaoo one enouoem an nonvone one on neaasao pens nosoo one oanononaoo one opoaoem Am unosnoaabno on» na opnonon uses one unease“ no aoaann_aaeam AH anaeuuoonp oo>aooaom Bog “H Haoo Hoaoaoo oagfiam opnoanoufi>nu no momma oeona nu mnaueuooo enouohm na oaenoa>fionn an ooonoahooum on on hpnaepaoonb oo>ao9nom oopoaooam one onoaonoaan 33m Hepnoanoafienm a wands ofiaenhn Odpmum 21 perceived uncertainty was experienced by those subunits experiencing the greatest amount of change (Cells III and IV). "The difference in per- ceived uncertainty between static and dynamic environments is always significant regardless of whether the environment is simple or complex" (1972:325). This discussion generates the following questions which provide a base for the hypotheses offered in Chapter Three. 1) Duncan found a relationship between perceptions of complexity and change and those of uncertainty. However his sample was one of related subunits. He comp pared twenty-two subunits of three manufacturing organizations (ten sub- units). All but one of his manufacturing subunits experienced simple environments (Cells I and III), while all but two of the research and de- velopment subunits experienced complex environments (Cells II and IV). Would these same results occur when independent organizations are tested? 2) A factor which has received attention in the work of decision theorists but has received little or no attention by organisation theo- rists is the relationship of past experience to perceptions of uncer- tainty. Is there a relationship between past organisational performance and perceived uncertainty? These questions provide the foundation for the first four hypo- theses of this study and will be tested across a sample of twelve inde- pendent manufacturing firms. Summary A review of the literature shows that, although the environment is a critical factor in the theorist's treatment of organizational behavior, the concept of organizational environment has not yet been clearly alized ‘ strateg; ronment componec based or complex: 22 specified or identified. lost of the attempts to date have been organi- zationally specific (e.g., Duncan, 1972), or too vague to be operation- alized (e.g., Thompson, 1967; Lawrence and Lorsch, 1969). A model and strategy for identifying relevant components of an organization's envi- ronment which is not organization-specific is offered. Environmental components which are common to all organizations are identified, and, based on prior theoretical and empirical work, the dimensions of change, complexity and uncertainty are identified and discussed. CHAPTER TWO INTRA—ORGANIZATIONAL POWER From the time of Plato and Aristotle, through.lachiavelli's era, and up to the present, power has been a recurring central interest of mankind. slost discourses of the past have been highly normative: how to get power, or how to use it. Until the 30's power was more the con- cern of students of political and military processes. Since then, the study of power has become an increasing concern of the social scientist, mainly within the province of social psychology; Host of the attention has been at the individual and interpersonal level, e.g., great man theories, (Bales, et al., 1954); trait leadership, (611188111, 1963); group leadership, (Lewin, Lippit and White, 1939); and/or the vertical or hierarchical dimensions of power, e.g., Tannenbaum's work with con- trol graphs (1968), although sociologists turned their attention to‘ studies of community power in the 50's, (e.g., Dahrendorf, 1959). Host often, power is treated as an independent variable, whose characteristics affect the variable under study. Swanson (1967) used power as the key independent variable to explain the acceptance of various forms of Protestantism in Reformation Europe. Weber's (1947) analysis of power led to his famous discussion of bureaucracy. The Human Relations school uses power to explain differences in morale, aelfbactualization, initia- tive and productivity of workers (e.g., Likert, 1961). The more popular strategies of organization development focus on power equalization and 23 improve: power re I t dition" poser sy situatiol sole sue. or J0 De m from the tion, 'hj individm tione" (; A 8' the aitm The “ads 313 or t studies "ill be 24 improvements in organizational effectiveness are traced to changes in power relationships (e.g., Bennie, 1969). It has not been until recently, with the advent of the "newer tra- dition" and its conceptualization of organizations as decision making power systems, that power has been considered in terms of it being a situational characteristic of an organization. Identification of a vari- able such as power as a "situational" characteristic stems from the work oftJ. D. Thompson and assumptions he articulated in his book Organiza- tions In Action. His basic assumptions are that "human action emerges from the interaction of: 1) the individual who brings aspirations, standards, and knowledge or beliefs about causation; and 2) the situa- tion, which presents opportunities and constraints. Interaction of the individual and the situation is mediated by his perceptions or cogni- tions" (1967:lOl-102). A study may concentrate on one or more of the above: the individual, the situation, the interaction of both or some combination of the three. The understanding of human behavior will be developed through a synthe- sis of these three areas, arrived at through the contributions of studies which delineate and clarify the characteristics of each. Power will be treated here in terms of Thompson's second category - as a situa- tional characteristic. Hora specifically, it will be considered as a property of organizations and organizational subunits, rather than indi- viduals or interpersonal relationships. Consequently the unit of analy- sis will be organizations and subunits. It is common to speak of the "power of General Hotors" or the "power of the marketing department", but neither General Haters nor the marketing department is capable of exercising power. These terms are a reflect the be‘ lotors vork wi ful" un‘ those a and "was here. '1‘ Ecutives ality" ' base". 381: 3 coaliti is a bett‘ burtenors Participa Percent 1 York UDiV tion, bee and 5313: ahO‘ m1 be: of 1’s dergtandl 25 reflection of situational characteristics which are defined over time by the behaviors of individuals in those situations identified with General [store or the marketing department and which serve to shape the frame- work within which present behavior is expressed. A member of a "power- ful" unit will have alternatives available which are different from those available to a member of a "weaker" unit. The terms ”powerful" and "weaker" are properties of the respective units, and not the memp bars. This point was stressed by Stagner (1969) in his study of 217 axe ecutives in 109 companies. He found no evidence that a "strong person- ality" would ”win out” in opposition to a subunit with a "strong power base". Baldridge (1971), and Pfeffer and Salancik (1974) have argued that a coalition model of organizations as outlined by Cyert and.laroh (1963) is a better description of "reality" than other popular models such as bureaucratic or collegial. In the coalition model it is assumed that participants have divergent goals and values, and that power is an imp portant factor in determining outcomes. Baldridge, in his study of New York University, argues that power, rather than rationality or consulta- tion, beat explains the decision making behavior he studied. Pfeffer and Salancik, studying decision making at the University of Illinois, show that budget allocations are more strongly related to departmental power than to measures of departmental work load, national rank and num- ber of faculty. They emphasize the significance of subunit power in un- derstanding organizational decision making. While there is an acknowledgement of the importance of subunit power in the literature, the.§gggg of subunit power has received rela- tively little attention as evidenced by the small number of studies in th1s I have 4 Hining devel¢ organ: Hicks: were I (1974 in wh d171d inter. ence (Hick. 26 this area. A review of the literature discloses only two studies that have directly dealt with the bases of subunit power - the work of Hinings, et al., and Salancik and Pfeffer. The most ambitious work is that of an English group who have been deve10ping what they term "a 'strategic oontingencies' theory of intra- organizational power". The basis for their approach was discussed by Hickson, at al., (1971) and empirical results based on this discussion were later reported by Hickson, et al., (1972) and Hinings, et al., (1974). They conceptualize organizations "as inter-departmental systems in which a major task element is coping with uncertainty. The task is divided and alloted to the subsystems, the division of labor creating an interdependency among them. Imbalance of this reciprocal interdepend- ence (Thompson, 1967) among the parts gives rise to power relations" (Hickson, at al., 19713217). In a study of seven manufacturing firms, they tested the relationships between power and the variables of coping with uncertainty, immediacy of work flow, pervasiveness of work flow, and substitutability of subunit activities and found no single variable highly related to power by itself, however, taken together, all vari- ables were related to power in differing degrees with "coping first, then nonsubstitutability, and last pervasivaness" (Hinings, et al., 1974240). Salancik and Pfeffer (1974) explored the bases of departmental power at the University of Illinois and found departmental power to "be most highly correlated with the department's ability to obtain outside grants and contracts, with national prestige and the relative size of the graduate program.following closely in importance" (1974:453). Their conclusion is "subunit power accrues to those departments that are most instrum valued Th Follow: a task one anc created their E sivene: anotne: I111 e and to Bubum source taintl but a} tranal task ‘ teak : Cult 27 instrumental in bringing in or providing resources which are highly valued by the total organization" (1974:470). This study builds and expands upon the work of these researchers. Following Hickson, et a1., (1971) the organization is conceptualized as a task inter-related system where "subunits control contingencies for one another's activities and draw power from the dependencies thereby created" (1971:222). However, as Hinings, et al., (1974) point out, their study dealt with only one dimension of interdependence: perva- siveness. They did not deal with the effect one unit would have on another. Based on Hickson, et al., and Emerson's (1962) work this study will explore one of the possible relationships between subunits power and task interdependence. Salancik and Pfeffer's (1974) conclusion that subunit power is related to the unit's ability to bring in needed re- sources for the organization can be seen as a form of coping. Uncer- tainties can spring from both external conditions in the environment, but also from the nature of the task itself. Coping, in order to be translated into power, must be recognized. One measure of the amount of task uncertainty might be the perception of the degree of difficulty in task accomplishment. A subunit's recognized ability to master a diffi- cult task can also be seen as a form of coping. £2221: Power has been conceptualized as "potential acts" rather than as transactions actually occurring (Katz and Kahn, 1966:220; also: French and Raven, 1960:609: Parsons, 1967:308). Power is seen as a resource which can be used in a number of ways, one of which is to barter or ex- change between two parties in a relationship (e.g., Emerson, 1962; 28 Blau, 1967). This view of power as a capacity or ability to influence future events is one of the most common found in the literature (e.g. March, 1966; Olsen, 1970; Harsanyi, 1962; Lehman, 1969; Bierstedt, 1950; Blau, 1967). Power is generally related to individuals although some theorists include norms (Mechanic, 1967) roles, and groups, (French and Raven, 1960) and these individuals are usually located in some relationship with other individuals. Dahrendorf, however, focuses more on the indi- vidual. He sees power "essentially tied to the personality of individ- uals" (1959:166) and Mechanic prefers "force rather than relationship" (1967:197). Dominance or control is the central issue of power for many theo- rists. Dahl, for example, says: "A has power over B to the extent that he can get B to do something B would not otherwise do" (1957:202). This theme of power‘gzgg someone can be found, for example, in Blau: "control through negative sanctions", (1967:116); Skinner: "control over aver- sive stimuli" (1971:42), or Alderson: "Control over expectations" (1967: 574). Others in this vein are: larch (1955), Wrong (1968), Tannenbaum (1968), Hickson, et a1., (1971), Bennie, et al., (1958), Emerson (1962), and Harsanyi (1962). In these analyses, the "relevant frame of refer- ence" as Martin (1971:246) suggests, is that of the subordinate, for it is the limitations on his activity "which symbolize the existence of a power relation". Not all theorists focus on the subordinate. Weber's classic defin— ition is in terms of the power holder: "power is the probability that one actor within a social relationship will be in a position to carry out his own will despite resistance" (1947:152). Power, as defined here, dimens tionsh availak Kornha. ing re; men tr To here. tive an Bamplin P1811 ty Some of Pertine havfi to Specifi tions" Th that th depende POrtuni “Bed 88 29 here, could include dominance, but it allows consideration of another dimension - that of power as a degree of autonomy in an ongoing rela- tionship. In this view, power consists of the "range of alternatives" available while maintaining the system of relationships (Dubin, 1963:19; Kornhauser, 1966:215). This "freedom" from constraints while maintain- ing relationships may be what many organizational members have in mind when they use the term "power" (e.g., Jay, 1967; Perrow, 1970b). The problems of defining power are much too thorny to be discussed here. Every theorist mentioned above has a somewhat different perspec- tive and each is open to criticism. The discussion above is Just a sampling of the power literature, and does not do Justice to the comp plexity and depth of the subject. For an excellent in-depth analysis of some of the major themes, see Martin (1971). The advice of Dahl seems pertinent here: "The particular definition one chooses will evidently have to be made from considerations of the substance and objectives of a specific piece of research and not from general theoretical considera- tions" (1957:207). The framework chosen for the analysis of power in this study is that the exercise of power is done by individuals in a relationship of dependency. The exercise is mediated by the situation which provides op- portunities and constraints for the individuals. This model will be used as a guide in the following discussion of the power literature. Exchange theorists see power as arising out of unbalanced social exchanges (e.g., Blau, 1967). Power, in this view, results from the de— pendence of one party on another in order for the first party to obtain resources or services controlled by the second party. To Emerson (1962) power is the obverse of dependency. A has power over B to the degree resour' cited ; leg1t1: ency of neglect (1962) 1 scarce ( “89 his in? powe Counter Ho m“Vernal 13 left into con prompts ; 0r B's 1‘1 iscover an be F “W“ (1 me refill- 30 that B is dependent upon A in order to achieve his (B's) ends. Theorists have conceptualized power relationships in terms of the resources controlled by A - the "bases" of power. One of the most often cited is the typology of French and Haven (1960): reward, coercive, legitimate, referent and expert. A has power over B to the extent that: l) B desires what A controls (reward), 2) A is able to punish B (coerb cive), 3) B endows A with the right to exercise power (legitimate), 4) B desires to identify with A (referent) and 5) A has some special knowledge required by B (expert). In each of these cases, the depend- ency of B is assumed. This analysis, however, is incomplete because it neglects the strategies available to B to avoid dependencies. Emerson (1962) has suggested four options open to B: 1) he can forego the re- source controlled by A; 2) he can utilize alternate sources; 3) he can use his own resources; or, 4) he can coerce A into providing the resource. Any power relationship can be analyzed in terms of these power bases and counter strategies. However useful these typologies might be, they are not sufficient material with which to build a theory of power in social systems. What is left unanswered (or unasked) is the question of how it is that A came into control of resources desired by B, and further, what is it that prompts B to enter into a dependency with A. What is important is not A or B's responses to unique sets of stimuli, but rather the object is to discover patterns of responses that are stable over time. A theory of social systems must be built upon the identification of situations that will be predictive of the responses of a number of A's and B's. As Martin (1971) suggests, ongoing power structures may be "conceived as the result of a number of 'bargains' between the differentially depend. sanctii pendent in the presen: torica; opporti Presen: mien a fine th lu larch ( ; iSnorer): 0f relat fines t: have ide "indepen (Centers and "Bye latiOngh 31 dependent" which over time may "be used to provide a normative basis, to sanctify, future unbalanced exchanges" (19712254). Power relationships in the past determine to a large degree the relative differential in de— pendence and differential access to avoidance strategies, which shape present power relationships. Situations, then, which are in part his- torically determined, provide a matrix within which are contained both opportunities and constraints for individual exercise of power in the present. A distinction must therefore be made between relationships which are indicative of the exercise of power, and situations which de- fine the constraints shaping the potential character of relationships. Inch of the confusion in the literature of social power which led larch (1966370) to wonder if power was simply a term used to mask our ignorance may be attributed to the preoccupation with characteristics of relationships without due consideration of the situation which de- fines the potential nature of the relationship. For example, theorists have identified "intercursive" and "integral" power, (Wrong, 1968); "independent" and "incentive" power, (Harsanyi, 1962); "conjugal" power, (Centers, et a1., 1971); "veto" power, (Kornhauaer, 1966); "intermenber" and "systemic" power (Lehman, 1969); all of which are descriptive of re- lationships and not situations. The aim.of this study is not to investigate power relationships, but rather to attempt to begin an identification of relevant situational characteristics that provide the constraints and opportunities for the exercise of power in organizations. As mentioned earlier, power has historical antecedents. Past activities and relationships provide a precedent which is the root of, and a constraint on, present alterna- tives. A "reputation" represents a distillation of past history which comun; ledges: repute" arises. only 1‘ relsti ranges power, and res "any at reduce ‘ I‘m the sys “Piano Fails a Power, j (1961) , Wovide my Pure It tional 8 my Pow Vast ecu (891nm d, examp19, 3198mm, (1953:257 \ 32 communicates expectations about the present. A reputation for know- ledgeability in a specific area communicates the expectations that the reputation holder will be knowledgeable when a new problem in that area arises. Likewise, a reputation for power in a particular setting, not only reflects past exercises of power, but predisposes and shapes power relationships in the present. The reputation serves to define the ranges over which the reputation holder has the potential to exercise power, and while providing a range of opportunities, it also constrains and restricts the utilization of power. As French and Raven have noted "any attempt to utilize power outside the range of power will tend to reduce the power" (19603621). Ihen the setting is an organization, one of the characteristics of the system relevant in understanding the exercise of power will be the variance in reputation for power between subunits in the organization. While a reputation for power provides opportunity for expression of power, it also serves as a constraint on behavior. As Landsberger (1961) and Stagner (1969) have shown, organizational characteristics provide a limitation on the range of effective behaviors any individual may pursue. It is in the light of the above that criticisms of the "reputa- tional approach" to studies of power must be considered. In the comma, nity power literature, one of the major aims has been to identify rele- vant actors in power arenas - to learn "who got things done ..." (Spinrad, 1966:218). The results have not been consistent. Hills, for example, feund a "power elite" in American society (1956:244) while Rieaman, attempting to investigate the same issue, found "veto groups" (1953:257). Kornhauser (1966) traces this disparity to the flaws inhere. reputs‘ of the ant exe tions. mite i1 33 inherent in the "reputational approach". As discussed previously, a reputation is a distillation of past history, it merely is one component of the present situation, it does not address the dynamics of the pres- ent exercise of power which was the aim offllills and Riesman investiga- tions. Also, there is more fluidity in community power relationships than in organizations. Organizational subunits do not have a choice whether or not to interact with each other; they must maintain perma- nent relationships if there is to be an organization. An actor in a community power arena often not only has a choice of relationships, but also many times may choose whether or not to enter into them. Spinrad (1966) identifies two methodological approaches used in the study of power: reputation and event analysis. In the reputation approach, respondents are asked: "who is powerful around here?" From these responses, a large sooiogram of the community is constructed. In event analysis, the question asked is: "who was influential in this specific instance of decision making?" Different results are found when these different approaches are used. For example, there is a debate in the book Power and Democracy in America (D'Antonio and Ehrlich, 1961) between Delbert Miller and Robert Dahl over the disparity in their find- ings. Miller, using the reputation approach, found a pyramidal power structure; Dahl, on the other hand, found a more pluralistic structure using event analysis. Differing results have also been feund in organizations by re- searchers using these approaches. Perrow, (1970b) in a study of twelve manufacturing firms asked his respondents to "rank the power" of sub- units in their organizations and found sales ranked the most powerful. Hinings, et a1., however, using event analysis in seventeen decision areas cospan points power - M discus; tem p; ground: has nO‘ 34 areas found production to have the most power in their sample of seven companies (1974:31). J. D. Thompson, in his critique of Perrow's study points out that.Perrow's approach results in a "tendency to conceive of power in several ways, which may or may not be consistent" (1970:90). Much of the confusion surrounding the "reputational" approaches discussed above has to do with the questions of what is meant by the term power, and who has power. Past studies may be criticized on the grounds that power was conceptualized in global terms. In most cases it has not been made clear whether the exercise of power, relationships, or situational characteristics were being discussed. Influence in Decision.lakigg One of the major difficulties associated with the concept of power is that while it may be theoretically conceptualized as the ability to win one's will over others, an operational identification must be based on the.gggg of power. Perrow (1970b) in his study of twelve firms asked organization members to "rank the power" of four organizational subunits. Upper management ranked sales the highest, while middle management gave this ranking to finance. Thompson (1970), in his critique of Perrow's article, suggests this might be attributed to differences in the re- spondents perceptions of the uses of power. .liddle management seems to define power in terms of the ability to affect the state of the system. This is not so surprising in view of the central interests of the dif- ferent levels of management. If it is assumed that upper management is most concerned with the system as a whole, then it seems reasonable that it should identify what Lehman calls the "systemic" dimension of power - the ability to "set, pursue, and implement goals for the system as a whole" agemen ends a manage. effort mentat. alloca‘ emphas; ceptior W he has bee 35 whole" (1969:455). If goal setting is a major concern for upper man- agement, middle management is concerned with the means to attain the ends specified by upper>management (Simon, 1957). Consequently, middle management identifies power in teams of the effect on its implementation efforts. Resource allocation is a significant factor affecting imple- mentation ability. Finance's exercise of power is felt through resource allocation. Therefore, it follows that the power of finance should be emphasized by middle management. The point to be made here is that per— ceptions of power will vary according to positions in the organization. Autonggy lost theorists, in discussing power have focused on relationships of asymmetrical dependency, and consequently, most of their attention has been on the conceptualization of power as influence over. However, when power is considered in an organizational setting of interdependent relationships, another dimension gains attention. Dubin (1963) and Kornhauser (1966) have equated power with the "range of alternatives" available to the power holder within the limits of maintaining the sys- tem. Crozier (1964) links power to "discretion". Downs (1967) sees power expressed in a bureaucracy in terms of subunit's attempts to main- tain and increase autonomy. To Jay (1967), power is freedom from con- straints. This is the ability of a party (subunit) in an ongoing rela- tionship (organization) to choose a variety of alternatives while still maintaining the relationship. In an organizational setting, this can be the degree to which a subunit can change its activities on its own. One of the difficulties with this as a measure of power is that it could be a function of interdependence. A subunit that was only peripherically 36 involved in the organization could have a large measure of autonomy but very little influence. However, under conditions of "reciprocal inter- dependence" (Thompson, 1967:54) where each subunit involved poses con- tingencies for the other units, then a measure of autonomy could be an indicator of power. Interdependence One of the more common themes in the literature of power is that power is related to dependency. In any given relationship, A-B, it is assumed that A has power over B to the extent that B is dependent on A. Emerson (1962), as noted earlier, takes the position that power is the "obverse” of dependency. Most theoretical discussions (e.g., Bierstedt, 1950; Dahl, 1975: French and Raven, 1960; Emerson, 1962; Martin, 1971) focus on interpersonal relationships that are asymmetrical in nature. That is, the relationships discussed are ones of asymmetrical dependency, or vertical power relationships. This focus on vertical relationships is the one most generally followed in studies of power in organizational settings (e.g., Etzioni, 1961; Warren, 1968; Peabody, 1962; Julian, 1966; Tannenbaum, 1968). There has been less attention given to horizontal power relation- ships under conditions of interdependence where the degree of asymmetry is not clearly identifiable. In an organization, one of the major types of interdependency has to do with the division of labor, or task inter- dependence. J. D. Thompson notes that there are differing types of intraorgani- zational task interdependence. He suggests three categories: pooled; sequential; and reciprocal (1967:54). Pooled interdependence describes 37 the situation where the actions of each part of the organization do not directly affect the actions of the other parts. Each part contributes to the whole, and each is supported by the whole, but there is no direct interaction between parts. An example of pooled interdependence would be academic departments or branch sales offices. Conditions of sequential interdependence exist when there is an order to the actions of the organization parts. The output of one part becomes the input of another part. For example, the output of a stamp- ing plant becomes the input for an assembly plant. A major character- istic of this type of interdependence is the asymmetry in the relation- ship. Reciprocal interdependence characterizes those situations where the outputs of each part becomes inputs of others and their outputs and in turn inputs for the first parts. Under these conditions all units involved are penetrated by all other units. The subunits of sales, manufacturing and engineering chosen for this study are examples of or- ganizational units in situations of reciprocal task interdependence. Blau and Scott (1962:183) also note the distinction between types of task interdependence. They identify two categories: parallel; and interdependent specialization, which correspond to Thompson's categories of pooled and reciprocal interdependence, respectively. It is important to differentiate between types of task interdependence as each of the types discussed above provides a differing Opportunity for the exercise of power, specifically along the dimensions of amount of interaction and degree of asymmetry. In situations of low interaction (pooled interde- pendence or parallel specialization), the opportunities for influence will be fewer than in conditions of reciprocal interdependence. While the question of this difference has not been empirically addressed, a 38 suggestion of this difference may be found by comparing two recent studies of intraorganizational power. Salanoik and Pfeffer (1974) in their study of university depart- mental power (pooled interdependence) found the question of resource allocation to be a major issue among the departments samples. Hinings, et al., (1974) in their sample of’manufacturing firms (reciprocal inter- dependence) found relatively less attention given to resource alloca- tion by the subunits and more to systemic or policy type decision areas. As discussed earlier, the identification of power must be based on its uses. Comparing these two studies suggests a difference in the uses of power which might be based in the differing types of interdependence in the organizations sampled. Consequently, it seems important in an in- vestigation of intro-organizational power to clearly identify the nature of the interdependencies of the subunits. There are two dhmensions of interdependence that are relevant for a discussion of power. They are: the amount of interaction character- istic of the interdependence; and the degree of asymmetry, or imbalance of dependency. A review of the literature shows only one attempt to investigate a relationship between subunit power and interdependence. Hininge, et al., (1974) found a relationship between "pervasiveness" and subunit power. Pervasiveness was Operationalized as the number of task interconnections of subunits, which is a measure of the amount of inter- action. As they point out in their summary, they did not investigate the degree to which the activities of one subunit are affected by the activities of another subunit (1974:40). As Hickson, et al., (1971) pointed out in an earlier paper, an imbalance, or asymmetry, in the degree to which subunits affect each other in task accomplishment 39 results in a subunit's ability to "draw power" from this asymmetry. Substitutability A major component in exchange theorist's conceptualizations of power-dependence relations is that of substitutability. The nature of the dependency is mediated by the availability of alternatives. One of the strategies suggested by Emerson (1962) for avoiding dependence is that of utilizing alternate resources. Blau (1967) makes the same point when he discusses monopoly as a base of power. Others of this theme are lechanic (1967), and Dubin (1963). Hickson, et a1., identify substi- tutability as a major variable in their "strategic contingencies" theory of intraorganizational power, which they define "as the ability of an or- ganization to obtain alternative performance for the activities of a sub-unit" (1971:221). Hinings, et al., in a later study based on this article, operationalized substitutability by determining "how easy or difficult it was to obtain personnel" as measured by: "level of formal education required for the job, length of experience and training re- quired, difficulties of hiring in the current labor market, existence of legal restrictions, number of tasks contracted out" (1974:26). A positive relationship was found between this variable and subunit power. The work of this English group is rich in providing suggestions for further research efforts. Their identification of substitutability as a base of subunit power is an important contribution. There are, however, alternate operationalizations that should be considered. If the idea is followed that "power is explained by variables that are elements of each subunit's task” (Hickson, et a1., 19713217), a corol- lary would be investigation of task substitutability. The English group 40 investigated one dimension of substitutability: that of personnel, they did not report any measures of task substitutability. Conceptually, these are two very separate dimensions of substitutability. It would be possible, for example, to replace all the personnel associated with ac- complishing a given task and have the task remain basically unchanged. Conversely, one could change the task and leave the personnel unchanged. The following discussion will attempt to identify this second dimension that of task substitutability. The first issue is whether the notion of direct task substituta- bility is a viable concept. Thompson (1967), for example, argues that the "core technology" is the essence of an organization and an organi- zation will go to great lengths to protect it from "environmental in- fluences". Substitution of a component of the technical core would re- sult in a radical alteration of the total organization, and the degree to which this substitution would be considered is questionable. If Thompson's assumptions are correct, it would seem that activities pe- ripheral to the core technology, and boundary-spanning activities would be more likely candidates for substitution. Another factor would be the immediate availability of activities outside the organization. For ex- ample, if existing market research or advertising services were avail- able, these activities would seem to be more readily considered for substitution. In practice, the notion of substitutability is probably more sub- tle. Direct substitution, or the threat of substitution does not seem to be a "normal" occurrence in most organizations. It seems reasonable to inquire into the strategies available that would preclude such con— siderations. Mechanic (1967) and others suggest some directions. One 41 strategy for maintaining nonsubstitutability would be control over in- formation regarding the subunit's activities. Activities that are not clearly understood will be harder to replace than those which are. Another possibility would be control over the appearance of difficulty in performing the task. Task activities which are seen as simple would seem to be more likely candidates for substitution than activities which are seen as more difficult. Two variables, then, that would contribute to substitutability would be the degree to which a subunit's activities are understood by others in the organization, and the degree of diffi- culty associated with the task. Task understanding and difficulty may also contribute to another concept associated with power - that of coping with uncertainty. Lawrence and Lorsch, in their study have included these two variables as components of uncertainty (1969:249). Thompson (1967) makes the point that uncertainties arise from both environments and technologies, or task activities. The mechanisms of coping have not yet been identified in the literature. One of the possible indicators of technological un- certainty might be the degree of difficulty associated with a given task. One of the difficulties with the concept of coping centers around the problem of identifying "successful" coping. If a subunit succeeds in "absorbing", "reducing", or "coping" with uncertainty, how is this recognized and differentiated from a situation that is inherently low in uncertainty? It seems that if coping is to be used as a base for power, the ceping must be recognized. The process by which this is done is not clear, but one requirement might be that there be a recognition of the difficulty associated with the task. 42 The Environment Coping with uncertainties in the environment has been suggested by many theorists as a base of intraorganizational power (Crozier, 1964; Thompson, 1967; Goldner, 1970; Perrow, 1970b; Hickson, et al., 1971). However, neither the process of coping nor the concept of uncertainty has been clearly identified. In fact, as discussed in Chapter One, there is, as yet, no agreement on the identification of the environment in the literature. Before one can investigate the relationships between power and the environment, the nature of the variables must be identi- fied and defined. An attempt to clarify this situation provides the underlying rationale for Chapter One. In that chapter, three dimensions of the environment were identified; complexity, change, and uncertainty. There is some disagreement in the literature as to the relationship be- tween subunit power and environmental uncertainty. The only study found which contains data addressing this issue is that of Lawrence and Lorsch (1969). Hickson, et al., interpret this work as showing that "marketing had more influence than production in both containerbmanufacturing and food-processing firms, apparently because of its involvement in (uncer- tain) innovation and with customers" (1971:219). However, a review of the data shows that while marketing received a higher score on the meas- ure of departmental influence (1969:111 and 127) in these firms, it did not receive a higher score than production on the measures of environ- mental uncertainty in both industries (1969:91). An unresolved question then, is whether there is a relationship between environmental uncer- tainty and subunit power. 43 Performance Performance as an independent variable has received little atten- tion in the literature of organizations. Social psychologists have noted the "Pygmalion Effect" (Bosenthal, 1973); one version is that knowledge of past performance leads to differential treatment of the performer. High performance in the past has been shown to lead to greater autonomy and influence. Ferris (1975) reports the findings of a number of studies which show that perceptions of past performance af- fect present conditions. High performers were allowed more influence and autonomy by their supervisors than were low performers. Salancik and Pfeffer, in their study of departmental power in a university, found that "power derived from acquiring resources is used to obtain more re- sources, which in turn can be employed to produce more power - the rich get richer" (1974:470). They assert "subunit power will be based on en- vironmental contingencies or important resources only to the extent that such feedback and constraints are perceived and recognized" (1974:471). The point here is that past performance influences present perceptions, and consequently, present power configurations. If coping activities must be recognized to be a base for power, then past performance can be seen as evidence of success in coping. A high performing unit would be displaying evidence of successful coping, and the power distribution would be expected to reflect this. If past performance is an important variable in explaining power relationships at a social psychological level, it may well be a salient factor in explaining subunit power. 44 The Critical Subunit The discussion so far has centered about the identification of power and its relationship to organizational variables. There is an- other issue related to subunit power that is found in the literature. Woodward has noted in her study of manufacturing firms that "there seemed to be one function that was central and critical in that it had the greatest effect on success and survival" (1965:126). Landsberger (1961) gives aneedotal evidence of the changing power positions of sub- units steming from changes in the critical issues facing the organiza- tion. In the firm he studied, during the depression years, financial control was seen as most critical and accounting was the most influen- tial subunit. This changed after the war when material shortages were common and the purchasing department eclipsed accounting in influence. Later, material shortages eased and the critical issue for the firm was identified as sales, and consequently the marketing department increased its influence. Perrow takes "the view that the most critical function tends to have the most power" (1970b: 66). These views seem to be based on impressionistic evidence. They have not been subjected to a test of empirical validity. A number of questions come to mind: is there, in fact, a relationship between the identification of a subunit as critical and the amount of power associated with that subunit? If so, what is the relation between criticalness and other bases of subunit power? Is the term "critical subunit" simply another term for power? Smary This chapter reviewed the literature on power and suggested that much of the confusion surrounding this concept is due to the fact that 45 in most cases, power is considered in global terms. It has not been clearly specified whether the exercise of power, relationships, or situ- ational characteristics were being discussed. Subunit power, the topic of this study, is a situational characteristic, and can be operational- ized by using event analysis. Subunit power has been suggested to arise from interdependencies of task activities and uncertainties stemming from both the task and the ex- ternal environment. These variables have been discussed and provide a ground for the remaining hypotheses to be offered in Chapter Three. Finally, the impressionistic evidence suggesting a relationship between perceptions of the criticalness of a subunit and the amount of power at- tributed to it has been noted and some questions raised about the nature of this relationship. CHAPTER THREE HYPOTHESES AND VARIABLES IN THE STUDY Introduction Chapter One discussed the concept of the environment and its place in organization theory. A model for environmental identification was offered and the environmental dimensions of change, complexity, and un— certainty were identified. The first three hypotheses discussed below will test the relationship between these three variables. A fourth variable - performance - has been suggested to be related to perceptions of uncertainty. This relationship will be tested by Hypothesis Pour. In Chapter Two, the concept of power, specifically subunit power, and its relationship to a number of organizational variables was dis- cussed. These suggested relationships provide the base for the remain- ing hypotheses. The final part of this chapter lists each variable in the study and discusses how each is operationalized. The Environment As discussed previously, the environment plays a central role in organization theory. One of the dimensions of the environment which has received much attention is that of uncertainty. J. D. Thompson, for ex- ample, argues that environmental uncertainty is a major factor in ex- plaining organizational behavior. He states: "technologies and 46 47 environments are major sources of uncertainty for organizations, and that differences in those dimensions will result in differences in or- ganizations," further, "the central problem for complex organizations is one of coping with uncertainty" (1967:13). However, it has been noted that, neither the components of the environment, nor the dimension of un- certainty has been clearly specified. The most advanced work in this area is that of Duncan (1972) who found a relationship between the envi- ronmental dimensions of change, complexity and uncertainty. Table 1 (Chapter One) represents his conceptualization of these relationships and provides a foundation for the first three hypotheses. Hypothesis One Units which are seen as operating in simple-static environments (Cell 1, Table 1) will experience the least perceived uncertainty. Hypothesis Two Units which are seen as operating in complexpdynamic environments (Cell 4, Table 1) will experience the greatest perceived environmental uncertainty. The assumption here, based on the work of Thompson (1967) and Udy (1959), is that changing and heterogeneous environments are associated ‘with increased uncertainty. Since it is assumed that a major component of'uncertainty is unpredictableness, environments which are simple in nature and relatively unchanging lend themselves more to predictableness than those environments which are complex and rapidly changing. Hypothesis Three Units which are seen as operating in simple-dynamic environments (Klell 3, Table I) will experience greater perceived environmental 48 uncertainty than units seen as Operating in complexpstatic environments (Cell 2, Table 1). The rationale is that the rate of change contributes more to uncer- tainty than does complexity. It is more difficult to predict and antici- pate the consequences of a situation that is changing because past prac- tices and procedures cannot be relied upon to be effective under the changed conditions. In a static environment, not as many unprecedented decisions have to be made and past procedures and practices will tend to serve as an effective precedent because of unchanging conditions. Uncertainty is related to the ability to identify possible outcomes and estimate probabilities of their occurrence. These abilities are based on experience. As Archer (1967) points out, the degree of uncer- tainty encountered is experience-related: the less experience in pre- vious situations of the nature being considered, the greater the uncer- tainty; the greater the experience base, the less the uncertainty. If we assume the central problem for organizations is "coping with uncertainty" (Thompson, 1967:13), or "removing equivocality from the in- formational environment (Weick, 1969:40), then we must also assume that 4am this behavior has taken place over time an experience base has been deveIOped on which present and future assessments of uncertainty are grounded. Estimates of future outcomes are "based on retrospective in- terpretations of actions already completed" (Weick, 1969:91). In other words, a "batting average" of successes and failures in prediction at- tempts is deveIOped over time and is a major canponent in the confidence level of new predictions. Present estimates of predictability have their roots in past successes or failures in prediction. Coping with uncertainty is a central element in the "newer tradi- tion" of organization theory. However, at this time, the process by 49 which this takes place has neither been clearly conceptualized nor operationalized. It can be assued, though, that if coping with uncer- tainty is a "central problem" for organizations, then one indicator of successful coping in the past would be some measure of organizational performance. An organization which has had a poor performance record could be expected to have a low confidence level in terms of predicting future events, and consequently, could be expected to identify its present environment as more uncertain than would an organization which has had a good performance record. In the sample of manufacturing firms in this study, one of the major measures of performance is economic. A firm that has experienced increasing sales and profits could be expected to seems that past caping has been successful, and consequently be more confident when assessing environmental uncertainty. Hypothesis Pour Low performing organizations will experience greater perceived en- vironmental uncertainty than will high performing organizations in simi- lar environments. The rationale for this hypothesis is that perceptions of uncertainty are grounded in experience. If performance has been effective in the past, the assumption will be made that competent predictions have been made in the past, contributing to a sense of confidence which will de- termine the confidence levels of present predictions, and consequently, perceptions of uncertainty. For the purposes of this study, the term environment will be taken to mean those factors, outside the boundaries of the organizations studied, which are identified as critically affecting the organization in the areas of labor, material, capital, product technology, process 50 technology, and markets. The method of investigation used is to ask re- spondents to list factors in the enviromment they feel are critical to the organization in these six categories. For each of the factors identified, respondents are asked to identify the amount of change and the degree of predictableness associated with each factor. The amount of change is identified as the static-dynamic dimension and is opera- tionalized by the variable of Environmental Change. The degree of pre- dictableness is called the uncertainty dimension and is operationalized by the variable of Environmental Uncertainty. Environmental complexity (the simple-complex dimension) is based on the number of environmental factors identified. Organizational performance will be measured by the actual amount of sales and profit increase over the past five years. .1292: The choice of dimensions of power to be used in a study is somewhat arbitrary and must be made in light of the researcher's interest. The aim of this study is not to investigate the exercise of power, but rather to begin an identification of relevant situational characteris- tics having a bearing on the exercise of power in organizations. One of ‘flhe dimensions of power chosen for this study is the degree of influence each subunit is seen as having in the specific decision areas of: 1) product innovation decisions; 2) marketing strategy decisions; and 3) capital budgeting decisions. The rationale for choosing influence in decision making as a dimension of power stems from the conceptualization of organizations as "decision making power systems" and the recognized effect subunit power has on decision making (e.g., Pfeffer and Salancik, 1971). These three decision areas were chosen because they represent 51 "systemic” (Lehman, 1969) issues. These are decision areas that Simon (1957) characterizes as "time-binding", that is, that class of decisions which set the ”rules of the game" or serve to limit and constrain later decisions. To demonstrate this point, he uses an example of a decision made to construct a shoe factory (1957:66). Once this decision is made, all subsequent decisions regarding product choice are influenced. The decision to produce shoes (in Simon's example) represents a sunk cost which restricts any alternative of later producing automobiles or any product other than shoes. In manufacturing firms such as those investi- gated in this study, decisions of this type have to do with: the nature of the product (product innovation); the method of marketing (marketing strategies); and the way the processes are financed (capital budgeting). Decisions made in these areas are time-binding in nature and affect later decisions. For example, making a decision to produce a standard- ized product at a large volume will affect the choice of production pro- cesses and sales efforts which would later affect the ability to produce a specialized, low volume product calling for different sales strategies and production processes. If, as Hickson, et al., (1971) assert, power {has a base in system task interdependencies, then decision areas affect- ing the nature of the task and its interdependencies will be important. ZPor this sample of firms, the important areas will be those dealing with product definition, marketing and financing. It is important to keep in mind that power should be considered in terms of specific issues, not general terms such as "most powerful." For example, Hinings, et al., (1974) report the production submit to ‘be most powerful in their sample of breweries and container manufacturers. While they do not report the raw power scores in their study for all of 52 the organizations in their sample, they do provide scores for one of the firms as an example of decision areas included in their power'measure. Three of their issues are similar to those chosen for this study. They are: marketing strategies, introduction of new products, and overall capital budget. when the reported scores on these issues are examined, it is found that sales is the most powerful, followed by manufacturing and than engineering. The point is that power is, to a large degree, the function of issues selected, and any interpretation of power studies must be made with this in mind. The second dimension of power chosen for this study will be a meas— ure of subunit autonomy. In Chapter Two, it was shown that most discus- sions of power assume relationships of asymmetrical dependency with the resulting conceptualization of power as influence over another. Under conditions of reciprocal interdependency (Thompson, 1967), such as are ganizations where subunits are not free to leave the situation, a meas- ure of power is the degree of autonomy a subunit possesses while still maintaining the organizational system. As the subunits chosen in this sample meet the conditions of reciprocal interdependence as identified by Thompson, autonomy, the degree to which a subunit can change activi- ties on its own, is chosen as a measure of power. In summary, two dimensions of power are chosen to be investigated in this study: influence in decision making in the areas of product in- novation, marketing strategies, and capital budgeting; and autonomy measured as the degree to which a subunit can unilaterally change its own activities. When the generalized term power is used in this study it will be taken to be comprised of the two dimensions just described. Power has often been related by theorists to dependencies. There 53 has been less attention given to situations of interdependence. In an organization, one of the major types of interdependency has to do with the division of labor, or task interdependence, which has been suggested by Hickson, et al., to be the "ultimate source of intra-organizational power" (1971:217). In a study reported by Hinings, et al., (1974), a relationship was found between the amount of task interconnections of subunits and subunit power. In their discussion they suggest some fur- ther directions that might be taken. The first of these is the degree to which the activities of one subunit are affected by the activities of another subunit. They suggest that the more a subunit affects others while accomplishing its own tasks, the more it poses contingencies for the other subunits, which provides a base for its power. A second sug- gestion is that the degree of imbalance, or asymmetry, in task inter- dependence contributes to subunit power differentials. This study will investigate the dimensions of interdependence sug- gested by this English group. While they suggested a direction for in- vestigation, they did not offer an operationalization of this variable. A step in this direction is suggested by the work of Lawrence and Lorsch and one of their measures of integration: the degree to which one sub- unit is influenced by the way other subunits perform their tasks (1969: 251). This measure will be used to construct the variables of task in- terdependence and asymmetry. Task interdependence is the degree to which a subunit affects other subunits while performing its own task, and asymmetry is the relative difference between the degree to which a subunit influences others while accomplishing its tasks and the degree to which it is influenced by other subunits. This provides the basis for the following two hypotheses. 54 Hypothesis Five The greater the degree of task interdependence of a subunit, the greater the power of that subunit. Hypothesis Six The greater the degree of asymmetry in task interdependence of a subunit in relationship to other subunits in that organization, the greater the power of that subunit relative to other subunits. Hypothesis Five addresses the question raised by Hickson, et al., (1971) in a simplified fashion. They pr0pose that to the degree that one unit's activities affect a second unit's activities, the activities of the first unit pose "contingencies" for the second unit, and there- fore become a potential base of power for the first unit. Hypothesis Six approaches the question of dependency from the per- spective of Emerson (1962) and others. Power differentials result from an hmbalance, or asymmetry, in the relationship. A subunit that affects others more by its activities than it is affected, portrays a situation where the other units are more dependent on the first unit than it on them. If power is the obverse of dependency, then the first unit should be more powerful. Hypothesis Five investigates the general question: is there a relationship between power and the degree to which the acti- vities of one unit affect other units? Hypothesis Six addresses the more particular question of asking if there is a relationship between imbalance and power. Another direction suggested by the literature has to do with the task itself. There are a number of themes that have been discussed. One of these is that power is related to substitutability. Honopoly of a needed service in a relationship provides a basis of power for the one holding the monopoly. If the service can be obtained elsewhere by 55 substitution, the power of the one offering the service is reduced. Hinings, et al., found a relationship between substitutability of sub- unit personnel and subunit power. They did not, however, deal with the issue of task substitutability. One of the difficulties with the notion of task substitutability is that, in an organization where the subunits are operating under conditions of reciprocal interdependence and whose activities constitute the core technology (Thompson, 1967) it is highly problematic whether task substitution on any scale is a viable consider— ation. In Blau's (1967) treatment of exchange theory, he discusses per- ception of task difficulty as a factor in social exchange at the inter- personal level. Appearing to master a difficult task can lead to in- creased esteem and power. However, to be seen as having difficulty with a task that is understood to be simple in nature, or easily accomplished has the opposite result. Blau's position is similar in many respects to those theorists who discuss the concept of "uncertainty absorbtion" (e.g., Crozier, 1964; March and Simon, 1958; Perrow, 1970b). "Absorbing", "reducing", or "coping" with uncertainty for a second party will provide a base of power for the first party. "Uncertainty itself does not give power: coping gives power" (Hickson, et al., 1971:219). It is not clear how either uncertainty, or coping is recognized. One possibility might be the degree of difficulty associated with the activity in ques- tion. This provides the rationale for Hypothesis Seven. Crozier, in his discussion of the bases of power of maintenance workers in the French factory he studied, suggests two factors relating to their power position. The first of these has to do with the issue of predictability, or uncertainty. In the factory, machine stoppages are 56 "the only major happenings that cannot be predicted" (1964:109). The uncertainties associated with such stoppages serve to provide a base of power for those responsible for remedying these stappages. He also dis- cusses a second, "complementary" factor - the lack of understanding as- sociated with the maintenance function. "No one can understand what they are doing" (1964:109). Because others in the organization are not able to understand the activities of the maintenance workers, Crozier argues that this also provides a base for their power position. As discussed earlier in Chapter Two, substitutability has been sug- gested by some organization theorists (e.g., Hickson, et a1., 1971) as being related to power. The level of understanding associated with a task could be a dimension of task substitutability. The less well un- derstood the task activities of a subunit, the more difficult it will be to identify a capable replacement for those activities. Conversely, controlling information so that activities are not understood by others has been identified as a strategy for influencing power relationships (e.g., lechanic, 1967). If lack of task understanding provides a base for power as has been suggested, then subunits whose activities are less well understood should have greater power than those whose activities are relatively better understood. Hypothesis Eight is intended to test this assumption. ‘Hypothesis Seven The greater the degree of difficulty associated with a task, the greater the power of the subunit associated with accomplishing that task. 57 Hypothesis Eight The less well understood the activities of a subunit are by others in the organization, the greater the power of that subunit. The preceding four hypotheses deal with the relationship between subunit power and "internal" organizational variables. We will turn now to a consideration of "external" or environmental variables. Uncertain— ties in the environment are an important factor in the theories of the "newer tradition". [any theorists have suggested a relationship between power and environmental uncertainty. The previous discussion on coping with uncertainty is equally applicable to environmental relationships. In Chapter One, three dimensions of the environment were identified: complexity, change and uncertainty. The relationship between these vari- ables and power is yet to be determined, consequently, they provide the basis for the following three exploratory hypotheses. Hypothesis Nine The greater the complexity of an environmental component, the greater the power of that subunit which deals with that component. Hypothesis Ten The greater the rate of change associated with an environmental component, the greater the power of that subunit which deals with that component. Hypothesis Eleven The greater the degree of uncertainty associated with an environ- mental component, the greater the power of that subunit which deals with that component. The preceding hypotheses have dealt with organizational variables. There is another issue having to do with subunit power which is referred to in the literature and is succinctly stated by Perrow: "the most 58 critical function tends to have the most power" (1970b:66). The as- sumption made here is that a subunit is powerful because it is critical. As Blau (1967) has pointed out, differentiation of power causes imbal- ances in social relationships which are a source of dissonance (Pestinger, 1957). This dissonance may be reduced by labeling the power differential "Just", "legitimate" or in the case of subunit power: "critical". Ihichever the causal relationship, the basic assumption that there is a relationship between a subunit identified as "critical” and its power has not yet been empirically tested. Hypothesis Twelve is designed to test this assumption. Hypothesis Twelve The more a subunit's activities are seen as critical, the greater the power that will be attributed to that subunit. This concludes the list of hypotheses in this studyi The following section discusses the variables and the manner in which they have been operationalized. Table 2 provides a summary of these variables. Table 2. Variables of the Study 1. The Environment 2. Simple-complex Environmental Dimension 3. Static-dynamic Environmental Dimension 4. Perceived Environmental Uncertainty 5. Subunit Power 6. Task Interdependence 7. Asymmetry 8. Task Difficulty 9. Task Understanding lO. Criticalness ll. Performance 59 The Environment The definition of the term "environment" as used in this study is based on the linkage model of organizetional/environmental interaction discussed in Chapter One and will mean those factors, outside the legal boundaries of the organizations studied, which respondents identify as critically affecting the organization in the area of labor, material, capital, product technology, process technology, and markets. Each of these six areas will be identified in this study as environmental £9!- ponent . For each of these environmental components respondents are asked to list the critical environmental factors of their own choosing which they feel affect their organization. For example, respondents are asked: "What do you consider to be the critical environmental factors (those outside your firm) affecting your firm's ability to obtain an adequate supply of labor?" The re- spondent then lists the factors he feels are critical to labor supply. The question is then repeated for each of the other five remaining en- vironmental components of material, capital, product technology, pro- cess technology, and markets (questions 12 through 17). The result is a list of environmental factors, arranged by components, which consti- tute the relevant task environment for that organization. For each or ganization sampled, the list of components will remain the same, howb ever, the factors will vary according to the respondent's perceptions. Simple-complex Environmental Dimension The simple-complex environmental dimension for each organization operationalized by the "Organization Measure of Environmental Complex- ity." For each of the six environmental components of labor, material is 9 choc gen; actu off '88 6O capital, product technology, process technology and markets, respondents choose a number of factors which they feel critically affect their or- ganization as discussed in the previous section. The number of factors actually chosen for each component by each respondent ranged from a high of five factors in one component to a low of zero where that component was not considered relevant by that respondent. An Organization Heasure of Environmental Complexity is developed by multiplying the total number of factors (F) identified by respondents in that organization by the number of components (C)2 identified. The re- sult is a measure of both the number of factors and the degree of simi— larity or dis-similarity of their effect on the organisation. The Index of Environmental Complexity is computed for each organization by (P) x (0)2. Squaring the number of components reflects the idea that variance in components affects an organization at more points than does variance in factors. For example, an organization with three factors in one comp ponent, and two in another would have an Environmental Complexity Index score of 20: (3+2)X(2)2=20. An organization with one factor in each of five components would have a score of 125: (l+l+l+l+l)X(5)2=125. A high score on the index indicates a higher degree of environmental com- plexity while a low score is indicative of less complexity. luch of the analysis of this study is at two levels: the organi- zation; and individual subunits. The Organization Heasure of Environ- mental Complexity is a reflection of that dimension for each organiza- tion. A different procedure is used to measure this dimension for sub- units. The three subunits in the study are: manufacturing: sales: and engineering. For each of these subunits, the relevant environmental 61 components are: process technology, market, and product technology, re- spectively. To compute a measure of environmental complexity for each of these subunits, the number of factors in that component related to the subunit in question as identified by all organization respondents is summed. The result is a measure of environmental complexity for that subunit. For example, if the respondents in an organization identified three factors in the environment which critically affected the technol- ogy used in producing the firm's products (process technology environ- mental component): two factors which affected the firm's ability to mar- ket its products (market environmental component); and no factors which critically affected the technology used in designing and developing the finm's products (product technology environmental component); the envi- ronmental complexity score for each subunit would be: manufacturing - 3, sales - 2, and engineering - O. Static-dynamic Environmental Dimension The static-dynamic dimension is a measure of the degree to which environmental factors are seen as remaining stable, or changing over time, and is Operationalized for each organization by the "Organization Measure of Environmental Change". For each of the environmental fac- tors that respondents chose, they were asked to select from a five-point scale of: (1) never: (2) almost never: (3) sometimes: (4) frequently: and (5) very often: the "number that best indicates how often each of the above factors (that had Just been named by the respondent) changes" (questions 12 through 17). As the units of analysis are organizations and subunits, not indi- viduals, the responses are pooled to reflect the situation experienced by ron eac EVE 62 by the organization as a whole. For the Organization Heasure of Envi- ronmental Change, this is done in three steps. First, a mean score for each respondent on each environmental component is computed. Second, an average of all respondent means for each environmental component is then computed, resulting in an organizational score for that environmental component. Finally, the total organization score is developed by sump ming the scores of each component. For a discussion of this pooling procedure, see Hage and Aiken (1967). Por subunits, the leasure of Environmental Change is simply the score for the relevant component as computed through step two described above. For both organizations and subunits, a high score on this meas- ure indicates more change, while a lower score indicates more environ- mental stability. Perceived Environmental Uncertainty The measure of environmental uncertainty used in this study re- flects the degree to which it is felt that an outcome can be success- fully predicted. For each of the environmental factors chosen by re- spondents they were asked to indicate how often they were able to pre- dict how changes in each of the factors they had chosen would affect their organization. The response categories varied along a five-point scale of: (1) never: (2) almost never; (3) sometimes; (4) frequently; and (5) very often, (questions 12 through 1?). Organizational scores on perceived environmental uncertainty are computed in the same three- step fashion as described in the section on Static-dynamic Environmental IDimension. Subunit environmental scores are also computed in the man- xier described above. The raw scores are reversed so that a high value ref; dice 63 reflects more perceived environmental uncertainty and a low score in- dicates less. Subunit Power Subunit power is measured along four dimensions: (1) ability of a subunit to change its activities on its own; (2) influence on product innovation decisions; (3) influence on capital budgeting decisions: and (4) influence on marketing strategy decisions. A measure of the first dimension - ability of a subunit to change its activities on its own - is determined by asking the respondents to indicate the degree to which each of the subunits, including his own, is able to change its activi- ties on its own (question 2). The response categories varied along a seven-point scale of: (1) not at all: (2) to a very slight extent: (3) to a slight extent; (4) to some extent; (5) to a great extent; (6) to a very great extent; and (7) completely. The power score for this dimension was computed for each subunit by averaging the pooled re- sponses to this question of all the respondents. For example, the power score for unit A on this dimension is computed by averaging the pooled responses of units A, B, C, and the president of the organization to the question of unit A's ability to define its own objectives. The same procedure is repeated for unit B and C, resulting in a power score on this dimension for each subunit. A subunit power score is computed in a similar fashion for each of the remaining dimensions of: influence on product innovation decisions (question 9); influence on capital budgeting decisions (question 10); and influence on marketing strategy decisions (question 11). An "Index of Subunit Power" for each subunit is then computed by 64 pooling the power scores of the four dimensions for that unit and de- termining the mean response score. The result then, is the Index of Subunit Power which reflects the perceived reputational power score over four dimensions for each subunit, with a high numerical value represent- ing more power and a low'numerical value representing less power of that subunit in that organization. Task Interdependence The approach used to determine measures of task interdependence used in this study is based on the work of Lawrence and Lorsch (1969:251). Respondents are asked to indicate the degree to which each of the sub- units, including his own, is influenced by how other subunits perform their own tasks. For example, respondents are asked to indicate the ex- tent to which the manufacturing subunit is influenced by the way the sales subunit performs its own task. When the respondent indicates his response to this question, he is then asked to indicate the extent to which the manufacturing subunit is influenced by the way the engineering subunit performs its own task (question 3). This procedure is repeated for each of the remaining subunits of engineering (question 4), and sales (question 5). The response categories varied along a seven-point scale of: (1) not at all: (2) to a very slight extent; (3) to a slight extent; (4) to some extent; (5) to a great extent; (6) to a very great extent: and (7) completely. The measure of task interdependence used in this study is a reflec- tion of the degree to which a subunit influences others while performing its own tasks and is computed in the following manner. For each sub- unit, the mean responses to the degree to which it influences other 65 subunits are summed to give a Task Interdependence score for that sub— unit. This procedure is repeated for each of the other subunits in that organization. For example, if the mean response score to the extent to which Sales influences lanufacturing while performing its own task is 5.25, and the mean response score on this same issue in relation to Engineering is 3.75, the Task Interdependence score for the Sales sub- unit in that organization is computed by summing each of the individual subunit Task Interdependency scores in that organization. A high value on this measure represents a high degree of interdependence, while a low numerical value reflects a low degree of interdependence. £12221 As discussed earlier, many theorists have suggested that an imbal- ance in task interdependencies is a basis for power differentials. This imbalance will be operationalized as the net difference in the ex- tent to which a subunit influences others while performing its own tasks, and the extent to which it is influenced by other subunits car- rying out their own tasks. This difference, or imbalance, will be called "Asymmetry," and will be computed in the following fashion. First, a value is computed for each subunit in a manner similar to that described above to compute the measure of Task Interdependence, which reflects the extent to which the subunit is influenced by the way the other subunits perform their own tasks. The mean responses to the ques- tion of the extent to which that subunit is influenced are stunned, and the result is subtracted from the Task Interdependence score for that subunit. The product of this computation will have either a positive znaaerical value signifying that that subunit influences others more by 66 its task activities than it is influenced; a value of zero signifying the interdependence is balanced; or a negative numerical value which signifies that that subunit is influenced more by other subunits in per- forming their task activities than it influences others. For example, if the computed Task Dependency score for a subunit is 9.00, and the score for the amount that subunit is influenced by others in their task activities is 9.50, the Asymmetry score for that subunit will be -.SO (9.00-9.50), which is a representation that that subunit is seen as be- ing influenced more by others than it influences other subunits in the area of task activities. What is at issue in this part of the investigation is to test whether there is any relationship at all between task interdependency imbalances and power. Consequently, consideration is given to the rela- tive differences in task interdependence scores, not the absolute values. It is possible that the relationship between imbalance and power differb entials is dependent upon the degree of task interdependence in an organ- ization. Task asymmetry might be more of a factor in explaining power differentials in organizations which are characterized by a relatively low overall degree of task interdependence than in organizations where there is a higher amount of task interdependence between subunits. Un- fortunately, the small sample size of this study precludes a considera- tion of this issue. Task Difficulty The measure of task difficulty chosen for this study is similar to that used by Lawrence and Lorsch (1969:249). Respondents are asked to indicate the degree of difficulty they see associated with three major tasks in the organization, (question 6) which are the task activities 85 ss.‘ 67 associated with each of the subunits of engineering, manufacturing and sales. These tasks are: 1) developing a product which can be manu- factured and sold profitably; 2) manufacturing economically a product which can be designed and sold; and 3) selling a product which can be developed and manufactured economically. Respondents are asked to choose a number on a seven-point scale ranging from a low of (1) no dif- ficulty at all, to a high of (7) extremely difficult, which reflects the perception of difficulty associated with each task. For each of the task areas a score is computed by averaging the pooled responses to that question. The result is a numerical value with a.higher number repre- senting a greater degree of difficulty associated with that task. The Task Difficulty score for each subunit is the value assigned to the activity associated with that subunit. For example, if the mean re- sponse value to the question of the degree of difficulty associated with deve10ping a product is 2.25, this value is the Task Difficulty score for the engineering subunit. This procedure is repeated for the sub- units of manufacturing and sales. The Organization Task Difficulty score is computed by summing the individual subunit scores on this dimension. Task Understanding Another variable related to task substitutability is the degree to which a subunit's activities are understood by others in the organiza- tion and is operationalized by a measure of Task Understanding which is determined by averaging the pooled reSponses to the following (question 7): respondents are asked to choose from a seven-point scale the ex- tent to which the activities of each of the subunits are understood, res i — — — zat SCO, der: pres 68 with the response category at the low end of the scale being "completely understood by those outside the department" and that at the high end be- ing "not at all understood by those outside the department." This is re- peated for all of the subunits in that organization. For each of the subunits, a Task Understanding score is computed by averaging the pooled responses to this question. A Task Understanding score for each organi— zation is computed by summing the individual subunit scores. A high score on this measure represents the perception of a low degree of un- derstanding by those outside the unit in question, while a low value re- presents a high degree of understanding. Criticalness To determine the degree to which a subunit is seen as "critical" relative to the other subunits in the organization, a measure of Criti- calness is determined in the following way for each subunit. Respondents are asked to rank the subunits in their organizations "in teams of the importance of each in contributing to your company's ability to compete successfully in your industry," (question 8) with 1 being the most critical, 2 being the second most critical, and 3 being the third. The rankings of all the respondents are pooled for each sub- unit with the resulting numerical value representing the overall ranking of each subunit. Performance The measures of performance used are those developed by Lawrence and Lorsch (1969:261). The presidents of each company were asked to indicate "the percentage change on a year to year basis of three 69 performance indicators: sales, before tax profits: and return on in- vestment before taxes" for the past five years. The result is a repre— sentation of the percentage change on each of these three measures for each of the last five years. In addition, each respondent was asked his subjective assessment of the entire organization as it relates to com- petitors in that industry. Summary This chapter has presented twelve hypotheses which will be tested in this study. Four of them deal with perceptual dimensions of the or- ganization's environment, while the remaining nine suggest relationships between subunit power and organizational and environmental variables. Eleven main variables of the study are listed and their operationali- zation is discussed. CHAPTER FOUR METHODOLOGY AND DEVELOPMENT OF THE INSTRUMENT Sample The main thrust of the study is toward a conceptualization of or- ganizational environments and an investigation of intra-organization power. As these areas have yet to be clearly defined in the literature, this study is of an exploratory nature, aimed at contributing to a more precise definition of these concepts. Consequently, the study is de- signed with an over-arching concern for simplicity, so that the rela- tionships between variables could be examined with a minimum of contamr ination from other sources. The following criteria guided the selection of organizations sampled in the study. 1) The organizations selected were to be the smallest, simplest, independent task systems that had distinct, identifiable sub- units in reciprocal interdependence (Thompson, 1967). 2) The organiza- tions were to be as similar in technologies, environments and size as possible. 3) There was to be a range of economic performance in the sample. 4) The functional divisions of labor were to be distinct and clearly the responsibility of identifiable subunits. 5) There was to be a relatively simple product line. To meet the above criteria, small manufacturing firms in the pri- vate sector were chosen. Further, there were two more reasons for choosing this class of organizations. First, there seemed to be a 70 71 sufficient number of these organizations in the local area that could be reached within the constraints of time and budget available for the study. Secondly, the researcher has extensive experience in this type of organization which contributed to a degree of rapport with the re- spondents. Firms were selected from the New England Manufacturers Directory on the basis of size and simplified product line. Sixty-three letters (see Appendix A) were sent out to the firms selected, followed two weeks later by a personal phone call to the firm's president from the re- searcher, asking fer participation in the study. At this point, twenty- eight firms expressed tentative interest in supporting the study. An appointment was then made with the president of each for an introductory interview during which the nature and intent of the study was explained in some detail and information was gathered as to the suitability of the firm as a research site. Fourteen of these were subsequently eliminated either because they did not wish to participate fully, or they did not meet the criteria listed above. This left fourteen firms which actually participated in the study. Because a number of firms agreed to partici- pate on the agreement that their anonymity would be assured, the firms are identified by code numbers - 101 through 114. Through the course of the study, two of the firms - 102 and 106 - because of changing condi- tions, decided they could not continue participation, and they were dropped from the final study. The final sample of twelve firms is listed in Table 3. Each of the twelve companies is autonomous, either independent of outside control (9 firms) or an independent subsidiary of a major corpor- ation (3 firms: 101, 105, and 113). The median size is 250 employees. 72 Table 3. Characteristics of Firms in the Study Company Identification Number of 1 Number Employees Technology Product 101 1200 B(v1) Hardware 103 150 B(VI) Electrical Controls Electra—Mechanical 104 250 B(v1) Measuring Devices 105 450 A(II) Ordinance 107 245 B(v1) Piano Keyboards 108 230 B(v1) Hand Tools 109 140 B(VI) Hand Tools 110 300 1(11) Blowers 111 200 A(IV) Bare Metal Plating 112 360 B(v1) Metal Fasteners 113 300 B(VI) Hardware 114 250 B(VI) Wire and Cable 1Technology is based on Woodward's typology (1958:11) A = Small Batch and Unit Production; B = Large Batch and Mass Production. Roman numerals in parentheses correspond to Woodward's sub-categories: (II) = Production of technically complex units; (IV) = Production of small batches; (VI) = Production of large batches, assembly line type. 73 They have similar technologies, nine of the twelve have "large batch" (Woodward, 1958) production technologies; all are located within the State of Connecticut, having in common a similar political and economic ”macro" environment (Osborn and Hunt, 1974). Organizational Subunits The subunits chosen for analysis in the study are those of menu! facturing sales and engineering. One of the difficulties encountered in the selection of firms was the wide divergence of organizational struc- ture. Of the fourteen firms who showed interest in the study, but who were subsequently not selected for analysis, most were rejected because the organizational structure did not meet the selection criteria, not because of a lack of interest in participating in the study. (A simi- lar range of organizational diversity is reported by Perrow, 1970b). Examples found were: the manufacturing function carried out in differ- ent locales, sometimes under the responsibility of semi-autonomous sub- sidiaries; more than one sales unit, or the sales function assigned to independent manufacturer's representatives, or some combination of the two; the engineering fanction integrated with sales or manufacturing. The firms finally selected for the study had one subunit each identifiable as manufacturing, sales, and engineering, although they did not share common labels. The distribution of functional responsibilities in the twelve firms is shown in Table 4. Data Collection Methods Data for this study were collected by means of a combined struc- tured and open-ended interview with the president of each company and 74 we: we: we: we: we: we: we: we: we: we: we: we: soaeoeeoam a. wow wow mam mew men men mum mam mam mum mew mew mowuewuepoeaamo easeoam .o and and we: use use and we: use we: we: we: 3.. ssoaoasooam 223-: .m we: we: we: we: we: we: we: we: we: we: we: we: awoaossooe doaooseoam .e moaam modem moamm moaem moamm ooasm modem modem modem moaom modem modem masseuse: .m OGNM omxm omflm OQHM OMNM OQNM OONM OflHfl omHm wan—OHSOOHA so so so so so so so so :2 5m can so H338 .m one son one one one one ssoaoasooam am am pom hem hem mm mm pom mm Mom on hem uonmq .H eds mas was Had oaa mod mod boa mos eoH mos Hos Assoc assaeoov HHZDMDm NHHAHmHmZOmmmm AdonBOZSN 333% no soaeadoasdoaoom assesses—fl v mnm