BD28 Dewey .H41A ALFRED P. WORKING PAPER SLOAN SCHOOL OF MANAGEMENT A STRUCTURAL EQUATION MODEL OF SCANNING BEHAVIOR OF MANAGERS Professor Sumantra Ghoshal Sloan School of Management August, 1985 W.P. #1696-35 MASSACHUSETTS INSTITUTE OF TECHNOLOGY 50 MEMORIAL DRIVE CAMBRIDGE, MASSACHUSETTS 02139 A STRUCTURAL EQUATION MODEL OF SCANNING BEHAVIOR OF MANAGERS Professor Sumantra Ghoshal Sloan School of Management August, 1985 W.P. //1696-35 A STRUCTURAL EQUATION MODEL OF SCANNING BEHAVIOR OF MANAGERS , Sumantra Ghoshal The Sloan School of Management Massachusetts Institute of Technology 50 Memorial Drive, Cambridge, MA 02138 June, 1985 First draft Comments welcome. Please do not cite, : The author is grateful for the comments and suggestions Westney, Donald Lessard, Michael Scott Eleanor of Morton, John Carroll, and Howard Stevenson; for help and in Seok Ki Kim and Yun Key Yun from assistance financial and for Korea; in research conducting field support from the Sloan School of Management, MIT. RiU.T. USft^^S SEP 2 4 1985 RECEIVED - page SECTION Policy Business called the models of organizations can of the that analytical through 1971), planning 1965; Ackoff, 1970), and through modifications Supporters emergent-incremental view, in contrast, have depended on empirical observations. from Quinn, do constrained strategy by ; rejected the view management" of incrementalist practice according the to navigating of limitations internal and adaptive along in and view, is an corridors the stream of disjointed decisions. as Believers actual in of of bargaining and coalition building that become manifest a that the "rational" process assumed external both process indifference They (e.g., Pettigrew, Strategic freedom of Organizations are formulation, incremental of follow not models. normative shown have 1980) organizations the claimed have They structures and processes (Chandler, 1962). deductions 1973; based their have adapt through managerial actions proactively Andrews, 1957; (Ansoff, their view and changing organizational purpose (Barnard, 1938; creating in followed In the strategy literature, behave. should have change. of how purposeful organizations models normative debate the to rational-analytical the on organizational and paths. parallel leaders) systems sides two the of Selznick, formulation strategy proponents of Organization Theory is between what can be and non-intersecting (or INTRODUCTION "rational-analytical" and the "emergent-incremental" Historically arguments : the major ongoing debates in the fields of both of One 1 - 1 rational-analytical view, labeling that strategic view as is both planning, and it as a "heroic unrealizable. in turn, have repudiated accurate nor useful. neither rebuttal from Andrews is typical unreal Thus, they have : the A recent . - page - 2 something to be said, I must admit, for the "There is owlish adherents to the primacy of incremental ism and administrative process. Repelled by the best and brightest analysts, they assert the wisdom of the organization, practitioner and believe that an properly nurtured, can lay out like a snail the path own So its progress. it can, but the path is of visible only behind the snail. Those who believe that life too uncertain to permit planning and that is must remain mostly intuitive, are a kind of purpose Greek chorus, keeping the rest of us honest while they hymn their classic cop-out." (Andrews, 1983, p3) differences manifest from extreme, the the as perspective in organization of field the In about diversity the role the similar theory, strategy of is views that range from, on one of choice argument of Child (1972) and, on strategic other, the population ecology model proposed by authors such Aldrich argued Astley by perspective In fact, as and Van de Ven (1983) and Pfeffer (1982), action on deterministic, Hannan and Freeman (1977). and (1979) or volun tarist ic the assumptions emergent or nature the regarding of decision- and choice is a critical dimension that distinguish among making different theoretical positions on the nature of organizations. Only debate this through Burgelman (1978), view contingency nature its of a part that versions such as Mintzberg of what can They called be In essence they argue strategy. freedom enjoyed by and is that entrepreneurship eclectic authors a that the and a firm depends on the nature of the characteristics of its own organization Dyer, 1983). The realized strategy of a firm has deliberate and an emergent component (Mintzberg, 1978); both part of and Lawrence and Dyer (1983). (1983), of environment (Lawrence work the organizational adaptation process and the extent the strategic conceptual synthesis is emerging out of a different proposed have of recently induced is by a the current structural context and a autonomous, (Burgelman, 1983). arising from managerial There is evidence that this and contingent view of strategy is receiving increasing - theoretical management page attention empirical and literature the in on (see Burgelman, 1985 for theory organization and - 3 a review ). view This There question. corporations. corporations, planning and argues, "autonomous". emerge strategy of then out of normative models. becomes the Yet, system is not fully the this process that most modern firms adopt ? : how does of "rational" In other words, through which, in the planning mechanism, process the is in the output the question The planning formal the world's largest of fairly extensive systems for long term have the "rationality" implied "induced" 500 that most of these systems are at least designed to Burgelman what over Klein and Linneman (1984) have shown than over 90% companies such as involving three-year, concluded recently their In study multiphased achieve that the increasing use of formal planning systems in large documents of an important literature empirical sufficient is begs however, strategy, of current strategy or structure of the firm affects its future the strategy ? Bower an had provided one answer to this question (1970) answer that focussed on the internal process through which alternative proposals are initiated and, problems are defined, finally, one of the proposals receives the "impetus" that leads to its Bower, Internal structure, in the process model of acceptance. affects strategy by influencing these sub-processes of definition, initiation and impetus. Environmental through which current scanning can be another mechanism strategy, despite an intendedly rational can circumscribe future strategy. Scanning is planning process, assumed to is the starting point for all normative strategic planning models. If, however, the provide current the external situation of intelligence the firm that determines what . - Pag e 4 - the environment will be paid attention to and how the of aspect Weick, by described making will be perceived (i.e., the enactment process sectors attended incremental the normative models, follows process or, if the actual even then 1969), strategy the output may be Burgelman's terms, display induced strategic in behavior scanning If called environment" "task the entirely is and limited to what Dill (1958) Thompson (1967) called the information base on which future strategy will be "domain", the based limited to the current activities of the organization. is cannot scanning Such strategy" "autonomous strategy input understanding what of must planning the that firm's the fields new in process domain is ; - to For autonomous must limited entirely not is by have an current i.e., a part of the task environment to the broader beyond look ventures to defined by Burgelman. as emerge, to information scanning lead "general environment". Therefore, behavior the factors An analysis of the following the preceding line of argument, provide can, insights into insights Such firms. to influence focus or breadth in scanning behavior of that interesting to which scanning is either limited extends beyond it ("breadth"). or managers issue in studying scanning environment (referred subsequently in this paper as current "focus"), interesting extent the is an strategy formation processes of the have can useful implications for both theory and practice. In the following individual level operat ionalized drawing for from reports paper This a explaining so the findings of such an analysis. section, the concepts of focus and breadth in scanning as to behavior are be measurable. diverse stream of literature, the differences in developed and In sections 3 and 4, a model is proposed this aspect of scanning. In - section - 5 data and the model are explained, and refining in methodology the 5, page section hypotheses, testing for and the findings are presented supports many of the theory derived data The 6. used also yields some results that are contradictory but to expectations. Limitations constrain generalizability the do provide conjectures. These findings sampling, of of basis the data, the for a data and results. analysis However, the of interesting number are discussed in the seventh implications and concluding section of the paper. Issues sampling of measurement and are briefly discussed in the Appendix. SECTION based is on "domain", as developed by Thompson a notion the (1958) of those as the "task-environment" parts of the broader are relevant to the setting and attainment of which environment of Dill by suggested OPERATIONALIZING SCANNING BEHAVIOR : concept The (1967) 2 Dill suggested that the task environment goals. Specifically, consists of groups. These, according to Thompson, constitute the membership of the domain notion the Defining (1966). organization-set is organization-sets organization generally, members the occupy" concepts the are "organization-set" developed by Evan there important the of are a focal organization's different statuses (Aldrich and Whetten, 1981). members organization's tied and regulatory subjective task and "there may be as many as to of of members a competitors A concept quite similar to that of the domain. firm's is suppliers, customers, together for an However, of the set are those who are task-domain for ultimately both through the notion of goals - - membership affect both of are page - determined organization's the 6 the actors who can by attainment goals of (Caplow, 1964). Following domain-constrained through obtained words, scanning related information distributors the current members customers, characterize of the domain. and and In other agents domain-focussed regulatory, social, and scanning information about general political, economic or resource-related through publications, consultants, advertising agents characterizes what may be called broad or unconstrained Such a scheme for categorizing scanning behavior scanning. is task -environment suppliers In contrast, obtaining etc. or about market, competition and behavior. issues focussed be defined as scanning for can information through technology to other acquiring arguments, these shown two types is made Figure in 1. Note that the division between the of scanning behavior is fuzzy and primarily in terms of what is the distinction the principal component of the manager's scanning activity. Residual broad environment Information Factors Immed iate task environment Within domain Outside domain Informat ion Sources Figure 1 : Categorization of scanning behavior - page 7 SECTION A paper by some suggest in reported in this and environmental on scanning have already suggested business Research in the fields of policy and organization behavior of others that have so far not been considered number a research organizational studies psychology, cognitive the influencing variables. these of of individual, of Previous attributes. THE INFLUENCING VARIABLES scanning behavior of individuals may be affected that host a : premise starting was 3 - scanning studies. influencers possible positions the in diversity, characteristics included in the study as individual The management hierarchy, their experience with entrepreneurial their and scanning behavior of managers are their of (there is no orientation assumption of the attributes being independent). Aguilar and Schoderbeck demonstrated managers, affect findings (1973) scanning (1968), and Keen management is (1980), (1973). individual's their derived from Schroeder, to comprehend and and tolerate and deal with ambiguity. is Kefalas on individual to diversity may the suggestions and Driver and Strenfert cognitive complexity thus improving her ability scanning , Wide experience with diversity can an individual's level exposure enhance to Keegan (1967) The possibility that the past background of behavior Hogarth , Hambrick (1982), among others, have of particularly, and, of and influence the behavior. scanning Collings (1968) (1967), integrate diverse pieces of information entrepreneurial orientation The possibility of the affecting the way carried out arises out of the arguments of Stevenson (1983) and other researchers on entrepreneurial behavior. While individual behavior is affected by individual attributes, it is also shaped by context - by the these page 8 - - characteristics tasks, and of the environment, orientation the with dealing decision making. systems. On environments fewer required for require complex scanning not may Managers have may categories information They other hand, organizations facing complex the environments may clusters sectors of of the organization. non-complex important critically the nature of managerial scanning of individual or extensive need Theoretically, environment. the in arguments follow from proposals of Thompson (1967) and these Galbraith King (1975) the findings of Fahey and Empirically, (1977). Kefalas and Schoderbek (1973) provide at and least limited supporting evidence. expectation The characteristics from scanning and between link a behavior of that nature evidence empirical the of task managers arises of the task represents one source of variation in managerial information processing requirements Tushman, 1969; greater need reduction performance available needs it can, managers to for uncertainty (Tushman and Nadler, Scanning is one of the Ven and Ferry, 1980). mechanisms scanning uncertainty and therefore information of effective de and high face amounts and Van 1978; When tasks are non-routine or highly 1979). participants complex, 1950; Becker and Baloff, (Bavelas, for meeting information therefore, be expected that the nature of depend on the nature of tasks that would a manager has to carry out. effect The system and on proposed been of tested the more analytically is its provides an prospecting depend more input firm on a for (Miles strategy on its scanning behavior of its managers have by Hambrick (1982). Put simply, firm is oriented, the more extensive system scanning firm's scanning the and a expected carrying and to out be such since scanning analysis. A Snow, 1978), in contrast, may experimentation and trial and error and may show lesser scanning intensity. - page These included be scan detailed discussion inclusion in (see Ghoshal, 1984 for In . the final and not the other variables for these of variables) such of a the framework is a subjective decision based on keeping judgement best are that This is not an exhaustive list and environments their managers choice variables number of other factors that can affect how a analysis, - influencing the model. the in may there are 9 view the conflicting demands of in parsimony and completeness. SECTION 4 Figure model that focus of the attempts of a variations in breadth and scanning level influences due to direct and as different factors discussed in the The hypotheses underlying the model section. preceding diagrammatic representation of explain to individual mediated MODEL AND HYPOTHESES the is 2 : are discussed below. perceived monitor to environment, the words, other members contrast, of environmental with : actors such continuity - and the relative stability in the stability, say, due a within the domain and it predictable environment, the needs of in a the predictable are monitored and responded to. signals be served by monitoring existing may cause the instability cope system the domain. of a stable oligopoly, there is usually an a intelligence external a In a When there is such these if from arises is manager feel the need to linearity - a signalling suffices that is existence of effective In there a environment the issues. of range environment. task more does the often continuity to be, diverse a unpredictable more The In an of domain and unpredictable environment, in may uncertainty itself and a lie in the broader range of issues may need to be monitored to uncertainty. Hence, it can be hypothesized - page 10 - CO o o CO w > w Q O En o w o o o Ed w o Pm o PL, PS P page 11 - - HI When the environment is perceived as predictable, the be focussed. When will scanning managerial is perceived as unpredictable, scanning environment will be relatively broad and unconstrained. : Environmental uncertainties. processing uncertainty managers model, develop by Daft predictability managers as outcomes and discontinuities (1974). than analysis is used by causal relations between the be or because there exists Under anticipated. limited actors and hypotheses contrast, In when environment the issues need a to be monitored. This leads to the : less analytically oriented. perceived as is environment When the in analytical will be more heterogeneous, managers their approach. H3 : Further, very large 1981) and a domain a heterogeneous organizational it networks environment (Aldrich results in and Whetten, becomes increasingly difficult to either define parsimoniously is large number In an unpredictable environment, managers will be : such This inverse value. need for analysis goes up since heterogeneous , the H2 seen as noted by Kobrin et al (1980), Boulton et al (1982) and been of is An environment is perceived unknown of As argued between environmental unpredictability and analysis relationship Keegan is environment rather cannot that analysis circumstances, has are is However, perceptions of the either because unpredictable actions when (1984), cope with uncertainty. to processes environmental homogeneity are and experimentation unpredictable, Hence, in the managerial of need for the for analysis differently. the need Weick and be analytical orientation. an endogenous variable. to affect more the perceived will more information the of that is orientation an as environmental expected argument environment, the analytical considered integral organizations of in to is a An view response to environmental analysis enough for the term to have any , page 12 - - practical connotation specified set linked. At the limit, in the case of extreme heterogeneity of or to limit scanning attention to any actors of to which firm the is directly the environment, the entire environment becomes a part of firm's the perceived become environment. task Thus, as the environment is to be more and more heterogeneous, broad, scanning has to involving more and more environmental actors. In other words, H4 : In a heterogeneous environment, relatively broad and unconstrained. Greater greater processed by Also, amount managers variety. task they unstructured. This essential has to be environmental information processed by also perceive plausible is environment their their tasks to be for the multiplicity of considered be be to effective task (the cause of environmental heterogeneity) is an that performance with environment, associated with their perception of perceive heterogeneous, issues that be This finding can be extended to suggest that managers when the will as shown by Daft and Macintosh (1981), positively is cope to environmental of of information of manager the uncertainty. the amount the is heterogeneity the scanning need feature to of an for unstructured task (Perrow, 1967). Therefore H5 : When the environment is perceived to be heterogeneous, managers perceive their tasks to be relatively unstruc tured . A with management's Analysis causal from the to emphasis commitment useful is effect becomes and strong when on to there analysis is usually associated firm's the is some of events on outcomes. current understanding of the The further away one goes immediate task environment, discern such causal implications. risk-taking domain. the more difficult it Experimentation rather than analysis are the characteristics of page 13 - - who managers look in analytical high with Managers environment. opportunities for the broader orientation their scanning within the task environment since it is focus such domain-constrained information that is amenable to only formal analysis. Or, H6 : More analytically oriented managers are also more focussed in their scanning attention. Experience of diversity enrich the cognitive script with managers (Abelson, 1976) and enhance their capacities to deal to limit the perceptions of ambiguity and equivocality compelled in Managers more exposed to diversity feel less ambiguities. with (Daft and Macintosh, 1981; Kiesler and Sproull, tasks their 1982). They are more willing and able to see the ill-structured aspects of their tasks and to acknowledge rather than ignore the of issues and influences that must be considered in multiplicity Therefore, managerial judgement and decision-making. H7 : Greater the managers' experiences with diversity, more they will recognize the unstructured aspects of their tasks. the Following orientation can (1983), an individual's strategic Stevenson be categorized as entrepreneurial or Trustees have the characteristics of bureaucrats; trustee-like. they attempt to find standard operating procedures and rules and they see those their aspects amenable explore the environment. broad the as that job ambiguity. and equivocal and structured. relatively standardization or They ignore ill-structured are and They formalization. not avoid Their schemas do not require them to often weak cues from the broader Entrepreneurs, in contrast, typically explore the environment acknowledge tasks. of to complexity tasks the for opportunities ill-structured Two hypotheses follow : and aspects are of more willing to their managerial - page 14 - H8 More entrepreneurial managers perceive their tasks to relatively less structured compared to be managers who are less entrepreneurial. : H9 : More entrepreneurial managers display relatively broader scanning behavior. and, by extension, analysis act as symbols Information and signals as (Feldman more to managerial actions and perceptions 1981). The more analytical a manager, the creates an internal environment that suggests analysis she be March, and guide that basis the characteristic amenable for ill-structured of formal to decisions actions. and tasks emphasize analysis typical a they are not that is quantitative) analysis because of the (and Therefore, managers who and ambiguity inherent in them. variety But, those who see their tasks as relatively are more structured, or H10: More analytically oriented managers perceive their tasks to be relatively more structured. Higher position the hierarchy, organizational differentiation (Lawrence unstructured her junior are managers, involve Top managers relatively of are more tasks the Lorsch, (Wellinsky, that complexity a is and decisions greater consideration less manager a a and top occupies extent the internal of and 1967) more Compared to 1967). manager explicit in has implicit or far diverse range of issues and to take influences. more involved in long term decisions that are impacted environment of the firm. by factors outside the current task Hence, Hll: Higher up a manager in the organizational hierarchy, more unstructured her tasks are and appear to be. H12: Higher the position a manager occupies in the organizational hierarchy, broader is her scanning. - page 15 - Finally, tasks their diversity to of of cm domain a unstructured, information becomes it of than task is the to identify the actors a particular course of who focussed attention characterizes perceive the to a scanning their tasks to be Hence the hypothesis: The more unstructured a manager feels her tasks to be, the broader becomes her scanning. SECTION The scanning based it managers highly unstructured. : take broader attention to the overall a of part particular to The whole notion when since difficult rather environment H13 fuzzy is Therefore, action. behavior problems. unstructured need to form the essential are that the variety and is they that affect the outcome of can who greater the judgements qualitative decisions be, unstructured managers see more the in 5 : TESTING AND REFINING THE MODEL model behavior South was of Korea. 111 tested and managers Sampling, refined using data on in 16 data large companies collection and measurement procedures are explained in the Appendix. followed nested design, with a survey the In principle, individual respondents belonging to different companies in two However, one-way broad sectors (trading and manufacturing). variance tests on all the measured variables indicated that sector or company level of analysis was not very useful since neither could explain a significant part of the variance for responses were treated Hence, the most of the variables. For a independently at the individual level of analysis. detailed discussion on the level of analysis, see Ghoshal (1985). * - general linear model developed by Joreskog and The Sorbom number and endogenous variables, prevented of the measurement model and only allowed testing of Exploratory model. causal analysis factor of indicators was used to obtain composite measures of essentially are therefore, effect, latent variables the model. in In latent variables were treated as if the variables, measured were they sample size, given the relatively large exogenous of multiple what used for testing and refining the model. small the testing the was (1982) However, page 16 - ignoring thereby separate However, all individual measurement errors. analysis correlations among different indicators of each of of the variables, as well as reliability coefficients of the composite measures convergent validity details) indicated satisfactory level of the measures (see Ghoshal, 1985 for of V, computer program that supports the the Joreskog, offers linear model of unweighted least squares general estimator depends a . LISREL the and methods for and the a choice between maximum likelihood parameter estimation. The ML method on the assumption of multinomial normal distribution of the variables but has the advantage of readily providing * To build model using all the concerned measured the indicators of the eight variables (i.e., all different would require the inclusion of 32 latent variables) variables and, following the arguments of Lawley and Maxwell rules out option. sample size of 111 that (1971), a three of the measured variables Further, between two and loaded significantly on the factors that, through the process of exploratory factor analysis, defined the measures of Thus, the sample size limitation the latent variables. prevented construction of even a simplified measurement model including only the most significant measured variables for each of the latent variables. - page 17 errors and t-statistics for the estimates. standard does method - The ULS assume any distribution for the variables, not but does not generate either standard errors or t-statistics for the nor an overall statistic for goodness of estimates fit. Given that most of the variables in the model are either ordinal or are computed by addition of variables measured on ordinal scales, ULS is clearly the theoretically superior method. However, the difference between ULS and ML estimates of the parameters are usually small and so, in the analysis, the ML method was used for model testing and refinement but the final model was also estimated on ULS. Instead of simply testing the proposed model, the of an iterative process of model testing, modifying approach and retesting was adopted. the starting The original hypotheses provided for the refinement process. point Parameters estimates were small compared to their standard errors whose eliminated were process the resulting model was retested. refinement model of The was guided by the analysis of residuals of the variance-covariance matrix and multipliers of the normalized legrange and log likelihood function (what Joreskog and Sorbom call "modification indices"). problem The absence associated with this approach is that of appropriate control, the final model may in the be strongly influenced by chance sampling error in the data. One way sample to into control two for halves this possibility was to split the so that one half could be used for model building and the other half could be used to provide clean test case, of however, possibility . the the refined model (Bentler, 1980). small sample size ruled a In this out that page 18 - - control solution The only on model refinement capitalizing possibilities of stopping refinement the condition proposed correct Lawley by overcorrecting minimize to the This was done by on chance. as soon Saris, de Pijper, and Maxwell specification real as tight a p('/' 2 as )X).10 and Zegwaart that this stopping criterion, originally shown have (1979) so process reached. was available was to adopt errors model the (1971), is most likely to spurious for model the in without error sampling (reference in Anderson, 1984, p31). SECTION 5 process The : FINDINGS modification model of and reconfirmation yielded parameter estimates shown in Table Before desirable of run review to 0.985 an adjusted 0.037. of The of freedom). between The value 0.973 of and an RMS was 0.297 with DLS and 0.316 with of the f variances of 5.57 (7 degrees were negative, no The correlations modification suggested indicator residual for index by 2.18 was Joreskog further - less than the and Sorbom (1982) as model trimming. The plot had a slope very near one compared o to the 45 The ULS estimates were small, the largest being -0.297. 5.0 normalized is The ML run yielded coefficient was more than one. largest positive . it The total coefficient of determination equations None the of measure ML estimation also showed a ML. correlation values, overall goodness-of-f it goodness-of-f it structural for parameter an RMS residual of 0.036. and residual the the model yielded an adjusted goodness-of-f it index the of reviewing 1. line (there was some skewness at the top). - PARAMETER HYPOTHESIZED SIGN page 19 - ULS ESTIMATE - goodness-of-f i t index The since interpreted distribution. bound All can that LISREL cannot be directly follow not determination of Walker per any known statistical be said is that it sets an upper sets the Weber and "suggests goodness-of-f it in does it in the percentage of variance explained by the model (the on coefficient as page 20 - that lower bound). (1984), a the high However, measure of major portion of the variance the data is explained by the model as a whole" (pl8). Figure parameters. The the path structure with the estimated shows 3 effects total of the different variables on scanning behavior are summarized in Table 2. Total effect on individual scanning behavior (extent of domain focus) Explanatory variables 1. Perceived environmental predictability 2. Perceived environmental homogeneity - 0.078 3. Experience with diversity - 0.029 A. entrepreneurial orientation - 0.089 5. Analytical orientation 6. Perceived extent of structure in tasks Table 2 : because methodology 0.095 - 0.178 Total effect of influencing factors on managerial scanning behavior These caution 0.171 that findings of must, however, be interpreted with the limitations in the data and constrain validity and general izab i 1 i ty of - page 21 - page 22 - - the after - error all (coefficient and determination) percent even difficulty with highlighted by that it with the users structural with models may is One way. models not often number of a path-structures and different data equally well and the only the made on behalf of particular model a superior to another model that shows is worse fit a Not having such an alternative model (few data. this technique ever do), even that claim cannot be of the final model developed made for the relationships measurement Finally, other the equation very fit be can that that may fit the data equally well users of the methodology is that path-strengths claim (adjusted 97 variance and there can be the relationships significant different and either not reflect these relationships or may may show of models alternative that model explains anything between 32 the of goodness-of-f it) many there is the possibility of modeling First, results. error or may they process the be out of development. model of artifact an arising spurious, be may Besides, in this study. of sampling, caused by non-random selection of respondents. On certain data tests internal of subjected was possible positive side, however, are the results of the to and For to. external validity that the want space, of it is not discuss those tests in this paper (see Ghoshal, 1985 for details), but the broad outcomes deserve mention. validity of the data (Kidder, 1981) was Concurrent tested by checking characteristics scanning of example, associations managers and concentrate conditions, managers use for certain intuitively behavior of between obvious managers. For functional specialization of kinds of information they gather (R&D managers technology, on and so on) marketing and sources managers they use on market (finance bankers, planners use publications, etc.) were all strong and of the right kind. - - page 23 analysis Factor validity indicators, the as well as supported construct generally analysis correlation simple of the variables and their operational izat ion of . In variables and measures were those that have most cases been successfully the previous researchers, though in by used different environments and contexts. Finally, managers sampled attributes and with (1983). that scanning modes range of variables such as personal attributes as well as the The directions surveys of scanning behavior of past of managers conducted researchers such as by Keegan (1967), Collings (1968), Kefalas and (1967), Schoderbeck and a with and results American Aguilar used, of these associations were largely consistent strengths North attributes such as kinds of of their environments and tasks. of the scanning behavior of the of sources associated were professional their terms in acquired, information adopted nature the Hambrick (1982) and Miller and Friessen (1973), provides some indirect (and weak) indication This the findings may be more generalizable than theoretical evaluation of the sample would allow. SECTION Out of the 7 : total six nature effects, predictability environmental perceived DISCUSSION tasks are strong. only and those of perceptions The effects of regarding the perceived environmental homogeneity, analytical orientation, experience are weak. with Four of diversity, of these expectations and two are not. and entrepreneurial orientation results are consistent with page 24 - - discussing In insignificant) possibly one's, strong the interpretation prove than weak the (and considered along with are spirit the the most generous that findings is that they suggest rather the hypotheses. of section, grounded as effects in set a introductory seen of results, these following the indicated As in the should be discussion and not as analysis of cause conjectures for demonstrated effects. When unpredictable, individuals themselves limiting scanning sociological (1967) The his in information processing structure for a for based (1967) was on his intelligence observed by Aguilar of organizations, is influenced As per the perceived by the review). Therefore, at least to a limited extent, how can the managers of a firm feel the environment to manipulated be at the margin through organization This perception of unpredictability, in turn, would design. affect need the organization (see Daft and Macintosh, 1981 of unpredictable be view unpredictability environmental This link the organization design. for not The association may have an studies. implication important and organizational the phenomenon same sector as ideological intelligence" Wellinsky of case broader a their task domain. for "political and study function. scan to by perceived is unpredictability suggested been had only environmental between broad environment the nature the of scanning and, in the long term, the adaptive capability of the firm. Norma t ively , For before. Scott Morton needs to it for broad both instance, such It an argument has been made has been suggested by Lorange, and Ghoshal (forthcoming) that an organization protect its core (Thompson, 1967) and expose effective environment learning and adaptation. is useful in Exposure to the sensitizing the organization . page 25 - - (or sub-unit) a isolating it Thus, according to these change. the structure of the organization or the sub-unit the vary depending on the particular phase of adaptation should that it to be overintegrated needs broaden to the wider environment. scanning their phase, change-embedding the the sensitizing phase so that at the environment to be more unpredictable perceive managers add An organization argument. normative this to findings present The through. going is credibility and protecting and ; desirable in the process of effecting and is institutionalizing authors, need for change the to needs it At be to overdif ferentiated so that members perceive the environment relatively more stable and can institutionalize the to be while paying attention only to the more limited task change environment dangers an overly analytical approach of companies that of Peters - sequence with of that scanning, be can unpredictability argument as domain-constrained is it to firm's outside promote discontinuity, . Such environmental of there task sensitivity may be a valid and to the broader their ability to innovate. enhance to opportunities technology era an members' the and its focussed argued, is incapable of finding bold In and and deemphasizing analysis and linear thinking so for environment A second, High analytical orientation leads to opportunities. new "ready-fire-aim" the weak, finding of the study is consistent argument. this follow to to overcome the hubris of analysis. and scanning need the admittedly though risk-taking. and Waterman (1982) have spoken about the "bias for and action" experimentation innovation, curb can of authors have recently pointed out the number A threats environment - A increasingly arise from particularly in the high industries where innovations in one industry can revolutionize another with which, till that time, it had page 26 - - little common. in unconstrained reactive model of gestalt a healthy dose of essential for proactive or even is events to their effects. discontinuity, causal the analytically, interpret to era, an Analysis is useful when there is connects possible complex a causal In the face may be too links intuition and and rather than logic and analysis may be the preferred learning of scanning coping. that such In under tools cannot with justified be gut-feel is managers will The first signs or of a threat is often a gut-feel that opportunity an such circumstances. analytical If such precision. felt to be acceptable within the company, not gradually stop looking beyond the immediate task domain. Managers constrained less are entrepreneurial two weak but implications Driver the Both have interesting placement managers of suggested had (1969) functions. staffing and should that match their functions in companies may require managers who are sensitive new These are the other complexities with the requirements of their jobs. cognitive Certain training the More orientation of managers, the findings. Streufert and scanning. their their scanning. expected for organizational is with diversity experience in personal the unconstrained more greater with to the broad environment strategic planning or development are possible examples. venture traditional more ; greater awareness compared to functions, of the accounting. systematically placing broader may environmental need a trends organization can benefit by An more marketing Even among entrepreneurial managers and those with greater exposure to diversity in such functions. Similarly, programs functions may to be internal training and development tailored so as to expose managers in such greater diversity by rotating them in page 27 - - assignments different new and companies have been following not are suggestions traditionally. them These locations. different in It comforting, however, that in is these implications, the findings are not counter-intuitive. All suggestions theoretical past What may be more in the surprises hypothesized, are two the as environmental instead of being more First, scanning, diverse restricted more becomes and Second, as the task is perceived to be less domain-bound. less structured, scanning becomes less diverse and more narrower focussed on diagram confirms the heterogeneity really one unstructured does lead an constrained et individual's cognitive maps. this hampered integrative the between complexity of Lawrence and Dyer (1984) have suggest that beyond to innovation and level of a learning Similarly, it can be argued that beyond "information complexity, task predominant are as means a avoidance" to a level behavior cognitive reduce Beyond this point, search becomes simple-minded overload. problemistic symptoms and demonstrated due to limitations in human information processing capacity. becomes finding complexity, information than less relationship U-shaped inverted complexity on have (1968) al environmental built rather ? Schroder of the surprise is therefore, perception of the task being the more to constrained scanning existence effect, In why : through the perception of mediated also is environmental of effect the path The environment. task that structure. task of review to model. increases, and empirical observations. and interesting heterogeneity as results and they confirm expected are these and alternatives" search ... (Cyert "... - in the in the and March, neighborhood of problem neighborhood 1963, pl21). of current In this . - page 28 - unstructured occupied respondents survey faced tasks. Therefore, it positions and judgemental and impossible not is All tasks. managerial unstructured relatively perhaps see decreasing scanning to increasingly with diversity expect indeed can we range, that the data pertains to this range. research by Ben-zur and Breznitz (1981) and Recent Streufert, stress of streufert the decision situation reduces the dimensionality in behavior. decision risk-taking Extending domain. implausible does not affect the extent of It leads to to a more unid imensional focus on but action one Denson (1983) also suggests that and when that argument, it is not this become too complex, decision tasks makers narrow rather than extend their scanning focus. Reviewing information search noted "... that benefits deficient, is even human in highly made skills a in and Serre (1983) have balancing the costs and acquisition be may seriously laboratory task in which the balance salient opportunities extended and of If these findings generalize to the are provided. learning Connolly tasks, information of on human performance in literature the real world information acquisition they imply significant non-optimalities noted earlier, tasks may be found in such tasks" (p3) also has It validity affects validity cues error and suggested information may Gilani, 1982). behavior. confirmed the show Edwards , in subjects information (1965) based Snapper suggestion that behavior. Low yield enough reduction in decision not normat ively decision perceived cue that acquisition to justify acquisition of such relationship which been on and (Connolly suggested this has the Bayesian model of Peterson (1971) have their laboratory experiments information behavior depart page 29 - - when information optimum from substantially d iagnost ic ity levels intermediate of information optimally information. Conolly finding show to information d : subjects is low, iagnosticity that under-acquisition diagnostic Serre (1983) have extended the the pattern of overacquisition of information of acquire and , at and decisions low-consequence for underacquire highly for only overacquire for and high-consequence robust to variations in overall cue validity. decisions is Further, they have that shown also such sub-optimal acquisition behavior persists with feedback and information learning is often limited. with of cognitive when stress ambiguity and environment Therefore the Given behavior. the information that accompany in from information the broader even lesser cue validity than information have more the they perceive their tasks to be equivocality tasks, unstructured managers face a high degree acquisition information their : This leads to an unidimensional focus unstructured. highly from scenario following the in of these findings is not inconsistent pattern The interpietable immediate task environment. unidimensional focus is directed toward that Hence the finding that more unstructured the task, domain. more domain-constrained the manager's scanning behavior. explanation for the finding may lie in the Another notion may slack. of have more may slack the resources capacity. can be invest to tasks in scanning the broader are highly unstructured, little be available for any activity other than getting immediate physical slack When environment. Managers' carrying out structured tasks slack but task done. in the The slack in question may not be sense of time or other tangible intellectual slack in the sense of cognitive Further, with structured tasks, most decisions taken through SOP's and the ones that need external - information environment environment the SOP's. may be because may implications from beyond the task from the task arising are consistent and standard and are included in Thus, the scanning needs for structured tasks biased toward task the environment information need since page 30 - is the is predictable broader environment precisely structured and and the immediate unambiguous and therefore does not need to be constantly monitored. - page 31 - APPENDIX SAMPLE June, In Republic to 16 agreed as in The well as questionnaire Telephonic participate in the study. to who those with follow-up firms 25 study the scanning practices of their managers. to firms interviews personal conduct largest the Korea (South Korea) were approached for access of surveys 1984, refused suggested that access firms perceived the research as an effort to find the those sources specific willing what with part to of their scanning and were not ways and they considered to be highly confidential information. contact was a Each company given 10 questionnaires and was requested to distribute managers approach selection management different in since adopted was respondents from of representation balanced ensuring internally, them This request, the and met personally. established was accepted that company each In a and levels of functions. proposal of random the list of company executives did not find favor with most of the participating firms. A were of total Given received. distributed firms filled and usable questionnaires 111 sampling the the process whereby the internally, questionnaires no evaluation of non-response bias was possible. cross-tabulation A level of the respondents by function presented in table 3. While all and management 111 respondents reported their management level (24 top, 26 middle, junior), 64 affiliation . is only 97 indicated their functional - page 32 - Functions . - categorized agents agents, publications, and consultants, general publications, trade shows, and others, trade were asked to estimate the percentage information external of distributors, and respondents the suppliers, bankers, advertising customers, as page 33 - they acquired from each of these sources competition and technology were considered Market, while environment, as and firm's task-domain, the immediate task resource, broad issues and that were part of the Customers, suppliers and were considered as members of the distributors agents to factors environment". "broad residual regulatory, considered were others relevant factors information as the other sources were considered and as representing outside-domain actors. Total time of each respondent for percentage of spent acquiring market, competitive, and technological information percentage of information gathered through and suppliers and agents and distributors were added customers, and scores sum was used as the a measure of the extent of domain- focus in the respondent's scanning behavior. 2. Environmental Characteristics rationale The that are perceived by managers as necessary for information (1962), decision making. Starbuck perception reality" and the hypothesis nature of the environment influences the kinds of the effective environmental including as variables of interest was characteristics that for of of (1976) and environment the its Following the arguments of Dill Weick (1969), it is the rather than any "objective characteristics that drives the attention interpretation processes collectively labelled as page 34 - behavior. scanning based managerial on for 1979, and not on any objective perceptions quantitative measures, Ireland, measurement of the environment were and opera t ional izat ion therefore, study, this In qualitative or discussion detailed a Downey (see and different on measures of environmental characteristics). Questions (1983) directly adopted dynamism, by proposed environment and along three authors the dimensions hostility, and heterogeneity - factor analysis of yielded only two distinct dimensions. interpretation of the Based loadings (see Table 4), factor dimensions were labelled as predictability (factor these and Miller However, while the study instrument. construct the responses the on in theoretical categorizes by for measuring environmental attributes were Friessen the developed scales and (factor homogeneity generated Note that these empirically almost perfectly match the environmental factors categorization 2). 1) scheme proposed by Lawrence and Lorsch (1967). Factor Predictability of competitors 0.75 Predictability of customer needs 0.61 Predictability of industry upswings and downswings 0.50 1 Factor 2 Diversity in methods of production and marketing 0.56 Rate of new product and process innovation 0.42 Table 4 : Measurement of Environment Attributes Factor Loadings . - 3. Task Characteristic s dimensions with percent of shown - suggested structured- and analyzability - factor only one significant of 1.34 and which explained 27 a categorization of tasks along unstructured dimension. dimension extracted two This extracted factor (loadings variance. the conceptualized had yielded eigenvalue an in Table 5) single variety responses the of authors the task of factor the while again, analysis the developed by Daft and Macintosh (1981). However, instrument here using measured were characteristics Task a page 35 - Here again, closely follows Perrow's proposal (1967) of categorizing tasks as rout ine-nonroutine Factor lot of experience 0.83 1. Tasks require and training 2. Tasks require extensive and demanding search for solutions a 3. Available information can be interpreted in several ways and can lead to different but acceptable dec isions Table 4. 5 : Task characteristics and 0.55 0.43 extracted factor Analytical Orientation Analytical a : 1 broader used questions orientation was one of the two dimensions of variable - strategic orientation - that was developed in and the study. scales The developed variable was measured using by Miller and Friessen (1983). . - The questions and are shown Table in variance). The significant any the results of factor analysis of responses 6 (the factors explained 41 percent of the first orientation. analytical page 36 - effect factor was used as indicator of an The second factor was found not to have on aspect the behavior scanning of reported in this paper and was not included in the model. Factor 1. Futurity of analysis 0.60 2. Top management involvement in analysis 0.41 3. Industry and competitive analysis 0.76 4. Use of analytical techniques Factor 1 0.64 5. Extent of risk taking 0.57 6. Aggressiveness in competition 0.66 7. Product/process/service innovation Table 6 : measured through Stevenson (1983). dimensions trustee-ship strategic resources, structure 0.50 Strategic orientation - underlying factors Entrepreneurial five 2 the The orientation concepts of individual proposed indicators and framework the was by proposed by him consists of for differentiating between entrepreneurial and orientation drive, control of commitment of managers. to resources, These opportunity, and dimensions are commitment preferred of management . page 37 - - All administered of the did had control and not respondents' an Stevenson by were However, only two questions, those measuring respondents' attitudes toward discriminating control proposed using five-point ordinal scales. commitment and questions the of resources, the on two a scale that was used as entrepreneurial orientation. their marginally but poor of commitment and dimensions of resources were added to form of Hence, the the tests of construct validity. pass strong The other questions were inter-correlated power. scores indicator showed acceptable reliability The scale (Cornbatch's alpha of 0.47) 6. of Experience with Diversity two Experience with indicators - countries lived in. diversity was operationalized as number of a sum languages read, and number of The resulting scale was found to have high reliability (alpha 0.76). 7. Management Level Management respondent based Managing top to on level was measured by assigning each one of three categories - top, middle, and junior their Directors, designations. and Executive Vice Presidents, Executive Directors were classified as managers, Directors and General Managers as middle managers, and Managers and Assistant Managers as junior managers. - page 38 - BIBLIOGRAPHY 1. Abelson, R.P., Decision Making", and Social Behavior Ackoff, 2. York, 1970. R.L., "Script Processing in Attitude Formation and J.S.Carroll and J.W.Payne (eds.), Co g nition Lawrence Erlbaum, Hillsdale, NJ , 1976. in . A Concept- el Corporate Planning, Wiley, New Aguilar, F.J., Scannin g the Business Environment New York, 1967. 3. Aldrich, H.E., Or g anizations and Environments Englewood Cliffs, 1979 4. . . McGraw-Hill, Prentice-Hall, H.E., and D.A.Whetten, Aldrich, "Organization-Sets, Making the Most of Simplicity", in Action-Sets, and Networks P.C.Nystrom and W.H.Starbuck (eds.), Handbook of Or g anizational Design Oxford University Press, Oxford, 1981. 5. : . R&D D.G., Mobility Innovation in and an PhD unpublished thesis, Graduate School of Business, Stanford University, 1984. 6. Anderson, Or g anization . Andrews, K.R., The Concept of Corporate Strategy Irwin, Illinois, 1971. 7. . Richard D. - "Corporate Strategy 8. The Essential Intangibles", at the Strategic Planning Conference of the paper presented Conference Board, New York, March, 1983. : Ansoff, H.I., Corporate Strategy An Analytical Approach to 9. Business Policy for Growth and Expansion McGraw-Hill, New York, : . 1965. 10. Astley, W.G., and A.H.Van de Ven Debates Organization Theory", in Quarterly 28, 1983. , "Central Perspectives and Science Administrative . 11. Barnard, C.I., The. Functions o_f University Press, Cambridge, Mass, 1938. the. Executive. Harvard 12. Bavelas,A., "Communication Patterns in Task -Oriented Groups", Journal of Accoustical Society of America 22, 1950. . Becker, S.W., and N.Baloff, "Organization Structure 13. Complex Problem Solving", Administrative Science Quarterly 1969. . and 14, - page 39 - Bentler, P.M., "Multivariate Analysis With Latent Variables Causal Modelling", Annual Review of Psycholo g y 1980. 14. : . N.R., Boulton, 15. "Strategic Planning Characteristics and Journal 25. 1982, : W.M.Lindsey, S.G.Franklin, and L.W.Rue, Determining the Impact of Environmental Academy Management Uncertainty", of . Bower, J.L., Managin g Resource Allocation Process 16. the Division of Research, Graduate School of Business Administration, Harvard University, Boston, Mass., 1970. . Burgelman, R.A., "A Model of the Interaction of Strategic 17. Behavior, Corporate Context and the Concept of Strategy", Academy of Management Review 18, 1983. . - "Strategy-Making and Evolutionary Theory 18. Toward A Capability-Based Perspective", Research Paper No. 755, Stanford University, June, 1984. : Caplow, T. , Principles of Or g anization 19. World, New York, 1964. . Harcourt, Brace and Chandler, A.D., Strategy and Chapters in the 20. Structure History of thz Industrial Enterprise. The MIT Press, Cambridge, Mass., 1962. : Child, 21. Performance J ,"Organizat ional Structure, Environment and The Role of Strategic Choice", Sociology 1972. 6, . : . Collings, R.F., Scanning the. Environment Strategic £&£ Information Unpublished Doctoral Dissertation, Harvard Business School, Boston, 1968. 22. . 23. Connolly, T., and N.Gilani, "Information Search in Judgement A Regression Model and Tasks Some Preliminary Findings", Organizational B ehavior and Human Performance 30, 1982. : . Connolly, T., and P.Serre, "Information Search in Judgement Tasks The Effects of Unequal Cue Validity and Cost", unpublished paper, Georgia Institute of Technology, May, 1983. 24. : 25. Cyert, R.M. and J.G.March, A Behavioral Theory of the Firm Prentice-Hall, Englewood Cliffs, 1963. . Daft, R.L., and N.B.Macintosh, "A Tentative Exploration into Amount and Equivocality of Information Processing in Organizational Work Units", Administrative Science Quarterly 26, 1981. 26. the . 27. as R.L., and K.E.Weick, "Toward a Model of Organizations Daft, Interpretation Systems", Academy of Management Review vol 9, 1984. , , . page 40 - - Dill, W.R., "Environment as an Influence on 28. Autonomy", Administrative Science Quarterly 2, 1958. Managerial , "The Impact of Environment on Organizational S.Mailack and E.Van Ness (eds.), Concepts and Prentice-Hall, Englewood Administrative Behavior Issues in Cliffs, NJ, 1962. 29. Development", in . R.D.Ireland, "Quantitative versus Downey, H.K., and Organizational in Qualitative Environmental Assessment Studies", Administrative Science Quarterly 24, 1979. 30. : . Edwards, W., "Optimal Strategies for Seeking Information", Journal of Mathematical Psychology. 2, 1965. 31. Toward Theory of "The Organization-Set a Evan, W., A pproaches J.D.Thompson (ed.), Interorganizational Relations", in Pittsburg Desi g n University of Press, to Or g anizational Pittsburg, 1966. 32. : . Fahey , L., W.R.King, and V.K.Narayan, "Environmental Scanning 33. Forecasting in Strategic Planning - The State of the Art", and Long Range Planning February, 1982. . 34. as Feldman, M.S., and J.G.March, " Information in Organizations and Symbol", Administrative Science Quarterly, 26, Signal 1981. J.R., Galbraith, Reading, MA, 1977. 35. Or g anizational Design . Add ison-Wesley Ghoshal, S., "Environmental Scanning State of the Art and a Research Agenda", unpublished paper, Sloan School of Management, MIT, 1984. 36. : An Individual and Environmental Sc anning Level Analysis unpublished PhD dissertation, Organizational Sloan School of Management, MIT, Cambridge, Mass, 1985. - 37 : , Hambrick, D.C., "Environmental Scanning and Organizational Strategy", Strate g ic Management Journal 3, 1982. 38. . J. Freeman, Hannan, M.T., and "The Population Ecology of Organizations", American Journal of Sociology 82, 1977. 39. . Hogarth, R.M., Judgement and Choice 40. Decision John Wiley, New York, 1980. : The Psychology of , Joreskog, K.G., and D.Sorbom, LISREL V Analysis of Linear Structural Relationships by Maximum Likelihood and Least Square Methods 1982. 41. ; . - page 41 - Keegan, W.J., Scanning the International Business Environment A Study of Lti£ Information Acqu isition Process. Unpublished Doctoral Dissertation, Harvard Business School, Boston, 1967. 42. : A study of Information "Multinational Scanning in Multinational Utilized by Headquarter Executives Sources Companies", Administrative Science Quarterly 19, 1974. - 43. : . Keen, P.G.W., The Implications of C ognitive Style t£LL unpublished doctoral dissertation, Individual Decision Making Harvard Business School, Boston, 1973. 44. . Kefalas, A.G., and P .P .Schoderbek , "Scanning the Business Some Empirical Results", Decision Sciences 4, 1973 Environment 45. . : Kidder, L.H., Research Methods Rinehart and Winston, New York, 1981. 46. in Social Relations . Holt, S., and L.Sproull, "Managerial Response to Changing 47. Kiesler, Perspectives on Problem Sensing from Social Environments Cognition", Administrative Science Quarterly . 27, 1982. : An Klien, H.E. and R.E.Linneman , "Environmental Assessment Internal Study of Corporate Practice", The Journal of Business Strategy, Spring, 1984. 48. : J.Basek, S. Blank, and J. La Palombra, "The S.J., 49. Kobrin, Assessment and Evaluation of Noneconomic Environments by American A Preliminary Report", Journal of International Business Firms Studies, 11, 1980. : D.N., and Lawley, 2nd. ed Statistical Method 50. . . , A.E.Maxwell, Factor Analysis Butterworth, London, 1971. Lawrence, P.R. and D.Dyer, Renewin g American Industry 51. Free Press, New York, 1983. as . a The Lawrence, P.R., and J.W.Lorsch, Or g anization and Environment 52. Richard D. Irwin, Illinois, 1967. Lorange, P., M.S.Scott Morton, and S.Ghoshal, 53. Control West Publishing, New York, forthcoming. . Strateg ic f Miles, 54. and Process 55. . R.E. and C.C.Snow, Or g anizational Strategy. Struct ure McGraw-Hill, New York, 1978. Miller, D. and P .H.Friessen , "Strategy Making and Environment Link", Strategic Management Journal, vol 4, 1983. - The Third Mintzberg, H., Science May, 1978. 56. . " Patterns in Strategy Formation", Management - page 42 - Perrow, C, "A Framework for the Comparative Analysis of Organizations", American Sociological Review. 32, 1967. 57. R. H. Waterman Peters, T.J., and Harper and Row, New York, 1982. 58. Pettigrew, A.M., The Politics Making Tavistock, London, 1973. 59. Search of Excellence In , of . Or g anizational Decision . . Pitman, D. Irwin, Pfeffer, J., Or g anizations and Or g anization Theory 60. Marshfield, Mass., 1982. Quinn, J.B., Strategies Homewood , Illinois, 1980. 61. for Change . Richard M.J. Driver, H.M. S.Streufert, Schroeder, and Information Processing John Wiley, New York, 1968. 62. Human , . Selznick, P, Leadership in Administration 63. Company, Illinois, 1957. Row, Peterson and . 64. Snapper, K.J., and C.R.Peterson, "Information Seeking and Data Diagnosticity", Journal o_f Experimental Psycholo g y 87, 1971. . and Their Environments", in Starbuck , W.H., "Organizations 65. M.D.Dunnette, ed . , Handbook oj Industrial and Organisational Psycholo g y Rand McNally, Chicago, 1976. . "A New Paradigm for Entrepreneurial paper, Harvard Business School, July, H.H., 66. Stevenson, Management", unpublished 1983. Thompson, 67. York, 1967. J.D., Organizations in Action . McGraw-Hill, New Sub-Unit M.L., "Work and Tushman, Characteristics 68. A Contingency Analysis", Administrative Communication Structure Science Quarterly, 24, 1979. : Tushman, M.L., and D.A.Nadler, "Information Processing as an Design", Academy of Organization Integrating Concept in Management Review 3, 1978. 69. . 70. Van D.L.Ferry, Measurin g and Assessing A.H., and Wiley, New York, 1980. Ven de Organizations f , 71. Weick, K.E., The Social Psychology of Or g anizing Wesley, Reading, MA, 1969. Wilensky, H.L., Or g anizational Intelli g ence 72. York, 1967. 1 2 5 k J 1 . . Addision Basic Books, New MIT LIBRARIES 3 TOfl D DD3 DbD 331 M^lMue 3K.0T** Lib-26-67