' Academy ol Management Review. 1986. Vol. U, No. 2. 280-297.

The Strategic Decision Process andOrganizational Structure

JAMES W. FREDRICKSONColumbia University

Contributions from the strategic decision process literature are synthe-sized and integrated with literature on organizational structure. Prop-ositions emerge that describe how the characteristics of an organiza-tion's strategic decision process are affected by its structure. Alsodiscussed are the patterns of strategic process characteristics that arelikely to be associated with different types of structures. Conclusionsare reached on issues such as the accuracy of alternative models ofthe strategic decision process, and the appropriate unit of analysisfor studying that process.

For many years authors have suggested thatthe relationship between organizational strategyand structure is reciprocal. Only recently hasthere been widespread agreement that structurecan have a profound impact on strategy throughits direct effect on the strategic decision-makingprocess (Bourgeois & Astley, 1979; Burgelman,1983; Fahey, 1981). A variety of strategic pro-cess and structural variables have been usedin describing isolated aspects of this relationship,and completing explanations have been pro-vided. However, most of this work remains frag-mented and major theoretical gaps persist.

This paper addresses the above problems bysynthesizing and integrating previous work; italso offers new explanations to fill critical gaps.This is a "first step" to encourage investigationand debate on how an organization's strategicdecision process is affected by its formal struc-

Donald C. Hambrick, R. T. Lenz, William H. Newman,Lloyd E. Sandelands, and Michael L. Tushman providedvaluable comments on an earlier version of this manuscript.

Requests for reprints should be sent to James W. Fred-rickson. Graduate School of Business, Columbia University,New York, NY 10027.

ture. It also attempts to encourage a broader de-bate, by assuming a perspective that is in sharpcontrast to both the "structure follows strategy"view and work regarding the effect that environ-ment and other variables may have on struc-ture. The paper begins with a review of litera-ture that traces the strategy/structure debate,identifies important characteristics of the stra-tegic decision process, and describes those di-mensions of structure that are most likely to affectstrategic decision making. The second sectiondraws on contributions from several areas todevelop propositions that describe how the char-acteristics of an organization's strategic decisionprocess are affected by individual dimensions ofstructure. In the final section, a link is establishedbetween each of the previously discussed dimen-sions and Mintzberg's (1979) well-known struc-tural "types." This makes it possible to describethe pattern of strategic process characteristicsthat is likely to occur in each type, and to under-stand why different structures are typically moresuccessful in different contexts.

The arguments presented here lead to severalconclusions. Among them are that the accuracyof alternative models of the strategic decision-

280

making process (Allison, 1971; Mintzberg, 1973),as well as the appropriate "unit of analysis" forunderstanding that process, vary with the typeof structure.

Strategic Process and Structure

Historical Relationship

In studying the development of America'sdominant industrial organizations. Chandler(1962) observed that major increases in unitvolume, geographic dispersion, and vertical andhorizontal integration were eventually followedby changes in structural form. Although a newstructure was adopted only after a change instrategy made its predecessor dysfunctional, thelogic of a relationship was compelling. In ad-dition, several early studies confirmed an asso-ciation between these variables (Fouraker &Stopford, 1968; Rumelt, 1974), and the proposi-tion that "structure follows strategy" becamewidely accepted.

In spite of the widespread acceptance of theabove relationship, there is a growing body ofliterature that suggests that there is a major effectfrom structure to strategy (i.e., once a structureis in place it will influence a firm's strategic deci-sion process, and ultimately its strategy). Forexample. Bower (1970) characterized structurebroadly, as the context within which decisionsare made, and observed that ". . .structure maymotivate or impede strategic activity. . ." (p. 67).Numerous other contributors (Bobbitt & Ford,1980; Duncan, 1979; Hedberg, Nystrom, & Star-buck, 1976; Jelinek, 1977) have argued simplythat structure constrains strategic choice.

To understand why it is logical for the strate-gic decision process to be affected by structure,one must understand the relationship betweendecision making and structure. March and Simon(1958) get to the heart of this relationship byarguing that an organization's structure imposes"boundaries of rationality" that accommodatemembers' cognitive limitations. By delimitingresponsibilities and communication channels,structure allows organizations to achieve "organi-

zationally rational outcomes" in spite of theirmembers' cognitive limitations (Simon, 1976). Italso helps management to control the decision-making environment and facilitate the process-ing of information. This link is apparent inBower's (1970) comment that "when managementchooses a particular organization form, it is pro-viding not only a framework for current opera-tions but also the channels along which strate-gic information will flow . . . " (p. 287).

Organizations often have some units whosestructures are different from that which charac-terizes the organization as a whole (e.g., whilethe organization is generally decentralized, someunits may be very centralized). However, thispaper is concerned with the structure that bestdescribes the whole organization, a concept thatwill be referred to as its "dominant" structure.Therefore, it is argued that the characteristics ofa firm's strategic decision-making process areaffected by its overall, dominant structure. Thiswill undoubtedly seem like an ill-founded asser-tion if one believes that (a) all strategic decisionsare made by one or a very few top-level execu-tives, or that (b) such decisions are made outsideof the dominant structure. As a result, it is impor-tant to illustrate that in many instances it is thesebeliefs that are ill-founded.

Regarding the first issue, it should be recog-nized that "choice" is only one of many activitiesthat are involved in the decision-making process(e.g., information search). Moreover, numerousauthors (Crozier, 1964; Mintzberg, 1979; Simon,1976; Thompson, 1967) agree that only in the mostsimple of organizations are all of the activitiescontrolled by one individual. Therefore, becauseit is difficult to obtain and comprehend all of theinformation that is needed to make strategic deci-sions in a large organization (Quinn, 1980), thestrategic process typically requires contributionsfrom people with a wide range of expertise andfrom numerous levels (Carter, 1971; Crozier,1964).

The second issue that warrants clarificationconcerns the strategic impact of a firm's domi-nant structure. It is recognized that organizations

281

may deal with strategic decisions by creatingtask forces, committees, and project teams(Thompson, 1967) and by using nominal group(Delbecq, Van de Ven, & Gustafson, 1975) orrelated techniques. However, the very nature ofthe strategic process makes it unlikely that suchmechanisms can fully negate the impact of thedominant structure. For example, Mintzberg(1979) has argued that a strategic issue can"emerge" from anywhere in an organization; itis not necessarily recognized first by those at thetop of the organization. Therefore, the issue maybecome evident to top-level managers, and maybecome the focus of a specially formed task forceor committee, only after it has been filteredthrough the organization's dominant structure.Moreover, it is suggested that the likelihood of afirm's using such mechanisms can be predictedby the characteristics of its dominant structure(e.g., a firm that is highly centralized is unlikelyto create special committees). As the first step inunderstanding precisely how a firm's strategicdecision process is affected by its structure, thenext section identifies several important charac-teristics of that process.

Strategic Process CharacteristicsMost studies of the strategic decision process

have produced either a very "focused" set ofobservations regarding one process question, ora very rich but "loose" description of the entiredecision process. An example of the latter is the"phases" and "routines" identified by Mintzberg,Raisinghani, and Theoret (1976). However, a re-cent comparison (Fredrickson, 1983) of the twotypes of models—the synoptic and incremental—that appear most frequently in the strategy for-mulation literature, identified several concept-ually distinct though related characteristics onwhich they differ. That comparison drew on thework of numerous authors, but was based mostdirectly on contributions by Lindblom (1959) andMintzberg (1973). In examining alternative de-scriptions of the strategic decision-making pro-cess, those authors reached very similar conclu-sions regarding the critical characteristics onwhich such processes could be differentiated.

The stream of work identified above concludedthat the dominant models of strategy formula-tion differ on the following six characteristics: (a)process initiation, (b) the role of goals, (c) themeans/ends relationship, (d) the explanation ofstrategic action, (e) the comprehensiveness ofdecision making, and (f) comprehensiveness inintegrating decisions. A description of these char-acteristics is provided in Table 1, which alsoidentifies critical questions regarding each. Al-though these six characteristics and their accom-panying questions are certainly not an exhaus-tive list, it is suggested that their basis in thetheoretical literature makes them particularlyimportant. Therefore, this paper will focus onhow they are affected by organizational struc-ture, following a brief discussion of that topic.

Structural Dimensions and Types

Structure refers to an organization's internalpattern of relationships, authority, and commu-nication (Thompson, 1967). It has been character-ized on a variety of dimensions and illustratedusing a variety of "types" (e.g., functional ordivisional). Moreover, debate continues regard-ing the validity of measures that have been usedto assess structure's dimensions (Blackburn, 1982;Fry, 1982; Walton, 1981), and the link betweenthe dimensions and types is often ignored. How-ever, three dimensions of structure—centrali-zation, formalization, and complexity—havereceived more attention than any others (Child,1974; Ford & Slocum, 1977; Fry, 1982; Hage &Aiken, 1967; Hall, 1977; Van de Ven, 1976) andthey appear to have the greatest implications forstrategic decision making. Each of these dimen-sions is also the dominant characteristic of a well-known structural type.

Centralization refers to the degree to whichthe right to make decisions and evaluate activi-ties is concentrated (Fry & Slocum, 1984; Hall,1977). A high level of centralization is the mostobvious way to coordinate organization decisionmaking, but it places significant cognitive de-mands on those managers who retain authority.

282

Table 1Critical Strategic Decision Process Characteristics

Process Characteristics

1. Process initiation

2. Role of goals

3. Means/endsrelationship

4. Explanation ofstrategic action

5. Comprehensiveness indecision making

6. Comprehensiveness inintegrating decisions

Description and Questions

Concerned with how and where the process is initiated. Is the process initiatedas a reaction to problems/crises, or the proactive pursuit of opportunities andinterests? At what level (operating or strategic) would a stimulus have to appearbefore members would recognize it as being strategic? What level assumes primaryresponsibility for initiating the process?

Issues regarding the role that goals play in the decision process. Will decisionsbe made to achieve individual versus organization-level goals? Will goals be"remedial" changes from the status quo, or "positive," future intended states? Arethe goals likely to be conceptualized in precise versus general terms?

Concerned with the relationship that exists between means (alternatives) and ends(goals). What is the likelihood that means will displace ends (goals) in thedecision process? Will goals persist in the face of significant changes in theavailable means?

Considers alternative explanations of the process that resulted in strategic action.Is strategic action most accurately characterized as intendedly rational strategicchoice, the result of standardized organizational processes, an internal processof political bargaining, or some other explanation? What is the likelihood thatstragetic moves will be incremental versus major departures from the existingstrategy?

Attempting to identify the factors that limit the comprehensiveness of thestrategic decision process. Is the primary constraint on the comprehensiveness ofthe strategic process top management's cognitive limitations, the detail achievedin the design of standardized organizational processes, or managers' parochialperceptions?

Concerned with how comprehensively individual decisions are integrated. Whatlevel of integration is achieved to form an overall strategy?

Mintzberg (1979) has discussed this issue by sug-gesting that an individual does not have the cog-nitive capacity or information that is needed tounderstand all the decisions that face a complexorganization. Therefore, it is not surprising thata negative relationship has been reported be-tween an organization's size and its degree ofcentralization (Pugh, Hickson, Hinings, & Turner,1968).

The degree of formalization specifies the extentto which an organization uses rules and proce-dures to prescribe behavior (Hage & Aiken, 1969;Hall, 1977). Therefore, formalization has signifi-

cant consequences for organizational membersbecause it specifies how, where, and by whomtasks are to be performed. A high level of formal-ization has the benefit of eliminating role ambi-guity, but it also limits members' decision-makingdiscretion. Therefore, it is generally argued thatthe level of formalization must be matched withthe level of professionalism because formaliza-tion threatens professional autonomy (Perrow,1972).

Complexity refers to the condition of beingcomposed of many, usually interrelated, parts.Regarding organizational structure. Hall (1977)

283

suggests that there are three pxjtential sources ofcomplexity—horizontal and vertical differentia-tion, and spatial dispersion. Therefore, an orga-nization that simultaneously has numerouslevels, broad spans of control, and multiple geo-graphic locations would be considered highlycomplex. While such a structure is often consid-ered appropriate for firms that compete in highlydifferentiated environments, it is important to rec-ognize that a high level of complexity makes itdifficult to coordinate and control decision activ-ities (Lawrence & Lorsch, 1967).

The above arguments suggest that the threedimensions of organizational structure havemajor implications for decision making. There-fore, in the following discussion each of theseconceptually independent dimensions—centrali-zation, formalization, and complexity—is dis-cussed in terms of its likely impact on the strate-gic process characteristics and questions iden-tified earlier in Table 1 (i.e., how and where theprocess is initiated, the role of goals, and so on).Table 2 summarizes the dimension-specific prop-ositions that emerge from that discussion.

Table 2Propositions Regarding the Effects of Three Dimensions of Structure

Centralization

Propxasitions l.A-D. As the levelof centralization increases, so doesthe probability that—

1-A. the strategic decisionprocess will be Initiated only bythe dominant few, and that it willbe the result of proactive,opportunity-seeking behavior;

1-B. the decision process will beoriented toward achieving"positive" goals (i.e. intendedfuture domains) that will persistin spite of significant changesin means;

1-C. strategic action will be theresult of intendedly rational,"strategic choice," and that moveswill be major departures from theexisting strategy; and

1-D. top management's cognitivelimitations will be the primaryconstraint on the comprehensivenessof the strategic process. Theintegration of decisions will berelatively high.

Formaiization

Propositions 2.A-D. As the levelof formalization increases, so doesthe probability that—

2-A. the strategic decisionprocess will be initiated onlyin response to problems or crisesthat appear in variables that aremonitored by the formal system;

2-B. decisions will be made toachieve precise, yet remedialgoals, and that means will dis-place ends (goals);

2-C. strategic action will be theresult of standardized organiza-tional processes, and that moveswill be incremental; and

2-D. the level of detail that isachieved in the standardizedorganizational processes will bethe primary constraint on thecomprehensiveness of the strategicdecision process. The integrationof decisions will be intermediate.

Complexity

Propositions 3.A-D. As the levelof complexity increases, so doesthe probability that—

3-A. members initially exposed tothe decision stimulus will notrecognize it as being strategic, orwill ignore it because of parochialpreferences;

3-B. a decision must satisfy alarge coristraint set, whichdecreases the likelihood thatdecisions will be made to achieveorganization-level goals;

3-C. strategic action will be theresult of an internal process ofpolitical bargaining, and that moveswill be incremental; and

3-D, biases induced by members'parochial jserceptions will be theprimary constraint on the compre-hensiveness of the strategic deci-sion process. In general, theintegration of decisions will below.

284

Effects of Structure's DimensionsImpoct of Centralization

As mentioned earlier, a centralized structureis one in which the right to make decisions andevaluate activities is concentrated. However,because one person seldom controls all strategicprocess activities even in a highly centralizedorganization (Mintzberg, 1979), in this section itis assumed that decision making rests with asmall coalition of top-level executives. Such aview is probably the one that most people haveof the strategic decision process, but few haveconsidered how structural centralization affectsthe characteristics of that process. The first col-umn of Table 2 indicates that the strategic impactof centralization is significant, and it first becomesapparent in its effect on how and where the pro-cess is initiated.

As is the case in any organization, membersthroughout a centralized firm are intermittentlyexposed to stimuli (e.g., problems and oppor-tunities) that have strategic implications for theirfirm. However, with such a structure strategicproblems and opportunities are likely to go un-recognized and ignored until they appear be-fore a coalition member. This is because Jcnow-ledge regarding the likely implications of indi-vidual stimuli, as well as decision-makingauthority, is concentrated in the hands of veryfew people. Therefore, although centralizationis a means of ensuring that decisions are tightlycoordinated, it may delay the start of the processor it may result in a complete failure to respondto some strategic stimuli.

Several authors have observed that strategicdecisions, in addition to being a reaction to stim-uli (problem and opportunity) can be proactivelyinitiated by the interests of coalition members.For example, the great "leaps" of Mintzberg's(1973) "entrepreneurial" mode represent the pro-active interests of a leader/founder who domi-nates a small firm. A similar phenomenon hasbeen observed in larger organizations by Carter(1971) and Gerwin and Tuggle (1978). The lat-ter authors suggest that the strategic process can

be initiated when the innovation interests of acritical mass of coalition members "coalesce" toyield proactive behavior. It should be empha-sized that coalition members may not be pro-active (e.g., some managers are very conserva-tive) and that the environment may not be recep-tive to such behavior. However, if they do indeedhave proactive interests, their interests are mosteasily pursued when those same members dom-inate the decision process. Therefore, it is sug-gested that centralization increases the likeli-hood that strategic decision making will be a pro-active, opportunity-seeking process.

As implied above, structural centralization alsocan affect the role that goals play in strategicdecision making. For example, under conditionsof centralization there are finite limits to theamount of diversity that can exist in the goals ofcoalition members. As a result, it is realistic tothink of members' goals as "intended futuredomains" (Thompson, 1967) that decisions areattempting to achieve. This is in sharp contrastto an organization that is not centralized, wherethe diverse preferences of individual membersserve not as goals to be achieved, but as "con-straints" on the decision process (Simon, 1964). Itis only when the dominant group is small, as it isunder conditions of centralization, that strategicdecisions are likely to be made with their "posi-tive" (i.e., intended future domains) preferencesin mind.

Simon (1976) has argued that all decisions havefact (means) and value (ends) elements, and thatin a rational model of decision making, ends areidentified before the means for achieving themare evaluated. In a centralized structure thismeans/ends relationship is closely related to theissue described above. Specifically, the goals ofcoalition members will guide strategic decisionmaking and they will have more impact on thestrategic process than under other structural con-ditions that are discussed later. Coalition mem-bers will also exhibit strong commitment to thegoals because they are their own (Latham &Yukl, 1975), even though they may be stated assimply "to survive" or "be number one." Because

285

of this fact, goals are likely to persist in the faceof significant changes in means, and the organi-zation may continue to pursue strategic goalsthat have been rendered obsolete by changes inthe means of competition.

It is widely acknowledged that a rational expla-nation of action, where an alternative is "chosen"based on its ability to achieve a desired goal,does not capture the reality of organizationaldecision making. Instead of goal "maximizing" itis characterized as a process of "satisficing," andinstead of being "rational" it is "intendedlyrational" (March & Simon, 1958). However, eventhough the decision process may not mirror arational model, a centralized structure gives top-level management an opportunity to make con-scious choices. Those choices are based on thepreferences of an individual or small group, butthey are directly reflected in the organization'smoves. Therefore, structural centralization facili-tates "strategic choice" (Child, 1972) and providesa modified (i.e., intendedly) "rational actor"(Allison, 1971) explanation of strategic action. Inaddition, because centralization makes it easierfor those who dominate to pursue any proactive,opportunity-seeking interests that they may have,it increases the likelihood that organizationalmoves will be major (e.g., Mintzberg's leaps), asopposed to incremental, departures from theexisting strategy.

It was suggested in the introduction that adecentralized structure accommodates members'cognitive limitations by factoring decision pro-cess responsibilities (March & Simon, 1958;Thompson, 1967). In contrast, a centralized struc-ture is uniquely susceptible to those limitations,and they affect how comprehensive the organi-zation is in making strategic decisions. Many con-tributions that question the ability of organiza-tions to make decisions comprehensively aregrounded in well-documented work with indivi-duals. For example, Steinbruner (1974) has ar-gued that the characteristics of his "cybernetic"model of organizational decision making are con-sistent with numerous cognitive theory princi-ples that describe a noncomprehensive decision

process. In addition, experiments like those ofBruner, Goodenough, and Austin (1956) supporttheoretical arguments regarding individuals' ten-dency to simplify decision situations (Braybrooke& Lindblom, 1970; Schwenk, 1984). Therefore, inan organization with a centralized structure, thecognitive limits of those few who dominate willdetermine how comprehensive the organizationis in making strategic decisions.

The final issue regarding centralization con-cerns its efiect on an organization's ability to com-prehensively integrate decisions to form an over-all strategy. Mintzberg (1979) points out thatelaborate structural configurations (e.g., matrixorganizations) are sometimes used to ensure thatstrategic decisions are integrated, but he arguesthat centralization offers a better solution. Withsuch a structure, the comprehensiveness of inte-gration is still subject to the cognitive limitationsof the dominant managers, but it may be higherthan if more formalized mechanisms are reliedon. For example, Quinn (1980) has suggestedthat strategic planning systems are helpful, butthat integration is generally accomplished onlyin the mind of the top executive. Mintzberg (1978)has provided support for this observation byarguing that a highly integrated "gestalt" strat-egy is likely only when the organization is con-trolled by a poweriul leader. Therefore, whilecognitive limits may restrict comprehensivenessin a centralized structure, they may have lessimpact on the process of integration.

Impact oi Formalization

Structural formalization is characterized by thepresence of rules and procedures that influencedecision-making behavior. As mentioned earlier,even when it exists only at low and intermediatelevels, formalization can affect an organization'sstrategic process as participxants gather and pro-cess information that is passed up the hierarchy(Carter, 1971). Moreover, Mintzberg (1979) hasobserved that firms that are highly formalized intheir "operating core" tend to be more formal-ized at all levels. In addition, the presence ofupper-level mechanisms such as budgeting sys-

286

tems are known to have an impact on the strate-gic process (Bower, 1970), and strategic plan-ning systems are a clear attempt to formalize de-cision making at even the highest levels. The se-cond column of Table 2 suggests that the impactof formalization can indeed be far-reaching.

Regarding issues of process initiation, it ap-pears that formalization increases the likelihoodthat the strategic process will be motivated byreactive (e.g., solving problems or crises), asopposed to proactive (e.g., searching for oppor-tunities), behavior. For example, Steinbruner(1974) has characterized organizational decisionmaking as a "servo-mechanism" whose pro-grammed responses are activated only whencritical variables get outside some specifiedrange. As is the case with Cyert and March's(1963) description of problems triggering stan-dard operating procedures, decision stimuli maybe ignored if they are not monitored by the for-mal system. In addition, it has been argued thatstrategic planning systems can become so for-malized that they drive out creative, proactivebehavior (Lenz & Lyles, 1983; White, Dittrich, &Lang, 1980). This suggests that a formalized struc-ture has the inherent ability to discourage thepursuit of opportunities.

A high level of formalization also affects therole that goals play in the strategic process. Forexample, by prescribing bounds of behavior, for-malized bureaucracies reduce goal incongrui-ties among members and provide reasonablywell-defined expectations about performanceevaluation (Ouchi, 1978). Because of this fact, itis expected that strategic decisions will be madewith precise, as opposed to general, goals inmind and that efficiency criteria will dominate(Simon, 1976). However, based on the aboveargument regarding the reactive nature of thestrategic process, under conditions of formaliza-tion these goals are more likely to be "remedial"corrections of the present state, and not "posi-tive, " future intended states.

The increased goal awareness that is broughton by formalization ultimately affects the rela-tionship between means and ends (i.e., the third

characteristic of interest). For example, formaliza-tion can produce what Merton (1940) has charac-terized as the "bureaucratic personality." In thisinstance prescribed behaviors become ends inthemselves, and means become more importantthan ends. This phenomenon helps explain whyformalized strategic planning processes some-times degenerate into exercises that produce lit-tle more than a bound document. However,Ouinn (1980) suggests that the problem cannotbe overcome by simply formalizing strategic-levelgoals, because doing so activates organizationalprocesses that are difficult to reverse.

Most strategic management literature explainsaction as being the result of a conscious choice.However, discussion in this section has pointedout that an organization that has a formalizedstructure is likely to respond to decision stimuliby employing standardized procedures (Cyert &March, 1963; Steinbruner, 1974). The variablesthat trigger the process are predetermined, andso are the possible responses. Therefore, strate-gic action in an organization with a formalizedstructure is most accurately characterized byAllison's (1971) "organization process" model. Itis the "outcome" of a limited cadre of capabilities.In addition, the actions themselves are likely tobe incremental. Ouinn (1980) has pointed out thatformalized strategic planrung processes tend toinstitutionalize incrementalism because they pro-duce actions that are only marginal departuresfrom the existing state. Similarly, the presence offormal monitoring mechanisms encourages suchorganizations to make incremental adjustmentsin response to feedback.

As implied above, the degree of structural for-malization will also affect how comprehensivean organization is in making individual strategicdecisions. Rules and procedures contribute tothe development of a firm's repertoire of behav-iors, and they dictate how various decision-making activities will be handled. For example,formalized search procedures increase the likeli-hood that information will be sought from areaspreviously utilized, and that solutions that weresuccessful in the past will be used again (Cyert

287

& March, 1963). Similarly, Carter (1971) hasdescribed how strategic alternatives are evalu-ated by applying "threshold level" analysis thatallows projects to be automatically accepted orrejected if they are above or below certain levelson a specified variable. Therefore, althoughformalized, "planned behavior" is instituted toachieve rationality in decision making (Simon,1976), the comprehensiveness of an organiza-tion's strategic process will be determined by thespectrum of behaviors that are accounted for inits rules and procedures (i.e., the level of detailachieved).

In addition to affecting how comprehensive or-ganizations are in making individual strategicdecisions, evidence suggests that formalizationcan also affect how well those decisions areintegrated. For example, Schendel and Hofer(1979) point out that formal policies have beenreplaced by formal planning systems as the pri-mary tool for trying to ensure that strategic deci-sions are comprehensively integrated. Althoughsuch systems may offer numerous benefits, theirability to achieve a high level of integration isagain determined by the detail of their design.Moreover, even the most elaborate planning sys-tem may not be able to achieve comprehensiveintegration because ". . .strategic decisions donot lend themselves to aggregation into a singlemassive decision matrix where all factors aretreated simultaneously in order to arrive at aholistic optimum" (Quinn, 1978, p. 17).

Impact of Complexity

An organization's structure offers three poten-tial sources of complexity: horizontal and verti-cal differentiation, and spatial dispersion. How-ever, the present description considers onlyhorizontal and vertical differentiation becausethey best illustrate the dilemma that structureposes as organizations try to accommodatemembers' cognitive limitations. Specifically,increased division of labor, which is manifest asincreased horizontal and vertical differentiation,requires increased coordination (Galbraith,1973). However, the pervasive effects of such

complexity do not become clear until each strate-gic process characteristic is considered indivi-dually. (These effects are summarized in the finalcolumn of Table 2.)

The impact of complexity first becomes appar-ent in its effect on how and where the strategicdecision process is initiated. Specifically, becausean organization's structure imposes boundariesof rationality on its members (Thompson, 1967),the degree of complexity specifies how wide orhow narrow those boundaries will be (i.e., ahighly complex structure has many, narrowlybounded positions). Therefore, since decisionstimuli (e.g., problems, opportunities) may maketheir initial appearance at any location in theorganization, the cognitive and motivational ori-entation that is induced by a particular structurewill affect how a stimulus is perceived and actedupon (Simon, 1974). Similarly, since strategicissues can emerge from anywhere in an organi-zation (Mintzberg, 1979), the degree of complex-ity will be a major determinant of whether mem-bers who are initially exposed to those issuesrecognize them as having strategic significance,or ignore them because of parochial perceptions.Therefore, if structurally imposed bounds arenarrow, as they are with a high level of com-plexity, members' self-interests may lead themto take no action, thereby leaving critical prob-lems and opportunities unattended or unex-ploited.

It is suggested that structural complexity alsohas an impact on the role that goals play in thestrategic decision process. For example, Law-rence and Lorsch (1967) reported that a high levelof complexity resulted in different goal orienta-tions across departments. This may explainBower's (1970) observation that different peopleinvolved in the strategic process are motivatedby different preferences. Therefore, in a com-plex organization the broad array of members'preferences or goals does indeed become aseries of "constraints" on the decision process(Simon, 1964), which makes it unlikely that strate-gic decision making can successfully achievesome specific future state. In addition, by restrict-

288

ing areas of responsibility and interest, a highlevel of complexity increases the salience ofindividuals' goals, and makes it increasingly dif-ficult for organization-level goals to influencedecision making.

Regarding the relationship between meansand ends, logic suggests that the level of com-plexity vill also have a strong influence on thischaracteristic. Task specialization accompaniesincreased complexity, and specialization fostersparochial perceptions. Therefore, it is expectedthat members in an organization that has a com-plex structure will have difficulty agreeing ongoals, and that the decision process will be itera-tive and political. Furthermore, because vastgoal differences may make it difficult to achieveconsensus on ends, managers may have to besatisfied with obtaining agreement on means,even though they accept the means for differentreasons. It is therefore suggested that the multi-ple effects of structural complexity will combineto produce strategic moves that are incremental,but for reasons (i.e., goal differences) that aredifferent than was the case with a formalizedstructure (i.e., institutionalization).

As implied above, conscious choice is also notlikely to be an accurate explanation of strategicaction in an organization with a complex struc-ture. First, because information may have to passthrough multiple organizational levels, the out-come of a strategic process can be affected(Carter, 1971). More importantly, horizontal andvertical differentiation not only create differingpreferences among organizational members,they also disperse power. These factors producea constraint set that is not likely to be satisfiedand that must be attended to sequentially (Cyert& March, 1963). Therefore, as argued by Petti-grew (1973), the division of labor that is manifestin a complex structure explains strategic actionas the result of an internal political process, adescription that is consistent with Allison's (1971)"bureaucratic politics" model.

As with the explanation of strategic action, theboundaries imposed by structural complexity alsohave an impact on how comprehensive an orga-

nization is in making individual strategic deci-sions. For example, Cyert and March (1963) haveargued that the "search" for decision-makinginformation is "biased" because participants'selective perceptions cause them to focus oninformation that is salient to the interests of theirdepartment or unit. As mentioned above, indi-vidual and unit biases are also introduced atmultiple levels as information is "preprocessed"on its way to the top (Carter, 1971). Therefore,the comprehensiveness of the strategic decision-making process will be affected by the extent towhich structural complexity evokes parochial(either individual or work unit) behavior fromparticipants.

The final issue regarding structural complex-ity concerns the extent to which this dimensionaffects an organization's ability to comprehen-sively integrate decisions to form an overallstrategy. Although organizational strategy is usu-ally characterized as a consciously integratedset of decisions, complexity creates problems forintegration. As discussed earlier, a complexstructure assigns a restricted range of decisionprocess activities (e.g., information gathering oranalysis) to members in a variety of locations(i.e., departments and levels). While this ad-dresses members' cognitive limitations, it alsoincreases the probability that actions taken inone unit will not be consistent with those inanother. Therefore, Mintzberg, (1978) has arguedthat a highly integrated gestalt strategy will onlybe common early in an organization's life, whenstructural complexity is low and power is cen-tralized.

It is hoped that the dimension-specific proposi-tions presented in this section seem importantand interesting. However, while these dimen-sions are frequently used in empirical research,practitioners and academics often think of struc-ture in terms of different "types" (e.g., functionalor divisional). Therefore, it could be both practi-cally and theoretically useful if the previous dis-cussion could be extended to describe how theoverall strategic process would look in organiza-tions that have different types of structures. Such

289

an extension would attempt to further illustratethe strategic process/structure relationship andpresent it in a context that has meaning for morereaders. Therefore, the arguments presentedabove provide the basis for the next section,which discusses the "patterns" of strategic pro-cess characteristics that are likely to developwhen different types of structures are used.

Patterns of Process Characteristics

Alternative Structures and Patterns

During the 1960s and 70s, several authors(Bums & Stalker, 1961; Chandler, 1962; Lawrence& Lorsch, 1967; Pugh, Hickson & Hinings, 1969;Rumelt, 1974) described alternative types ofstructures. As a result of their contributions, inves-tigators routinely refer to different structures withnames such as "organic" and "mechanistic,""functional" and "divisional," or "workflow bur-eaucracy," with the expectation that readers willhave a basic understanding of their characteris-tics and implications. Therefore, it would be help-ful if a link could be established between someof these well-known forms and the strategic pro-cess characteristics that were discussed in theprevious section. For example, how and where isthe process initiated? What role do goals play,and so on, in an organization with an organicstructure? How do these characteristics differ inan organization with a mechanistic structure?The empirical "archetypes" generated by Millerand Friesen (1977) represent one of the few at-tempts that have been made to establish a linkbetween structure and distinct patterns of strate-gic process characteristics, but they differen-tiated structures primarily on the centralizationdimension.

In addition to the structural types described bythe above authors, Mintzberg's (1979) synthesisof previous research (including those types) pro-duced five forms that have begun to appear inthe literature. The "Simple Structure," "MachineBureaucracy," and "Professional Bureaucracy"are the "purest" forms, while the "DivisionalForm" simply uses an over-arching administra-

tive structure to loosely link some combination ofthe other three. The "Adhocracy" is a large-scalematrix. Mintzberg also suggests that each struc-tural form results when one of several compet-ing "pulls" dominates the others. If this argu-ment is interpreted to mean that these structurestend to be dominated by one dimension, there isa basis for hypothesizing about the strategic pro-cess impact of different structural types.

It is suggested that Mintzberg's (1979) purestforms—the Simple Structure, Machine Bureau-cracy, and Professional Bureaucracy—are in factstructures whose dominant dimension is one ofthe three that were previously discussed. Morespecifically, centralization is the dominant dimen-sion in a Simple Structure, formalization domi-nates in a Machine Bureaucracy, while a Profes-sional Bureaucracy is characterized first andforemost by complexity. This link is illustrated inFigure 1, where the three types are mappedagainst the dimensions of structure. It is impor-tant to note that these structures are widely distri-buted across the matrix, which emphasizes thatthey are very different. Also, it is later illustratedthat these types are the most common amongorganizations.

The relationship that appears to exist betweenMintzberg's (1979) three types and the threedimensions of structure suggests that the aggre-gate of propositions that were previously attrib-uted to each dimension can be used to producea pattern of characteristics that describe how stra-tegic decisions are made in each type. In thefinal section each structural type is described, itspattern of strategic process characteristics isdiscussed, and the context where it can beexpected to be most successful is characterized.It is important to note that the discussion of pat-terns draws heavily on the arguments that weredeveloped in the previous section. Therefore,only the primary conclusions (and not the sup-porting literature) are provided.

Centralization and Simple Structure

Mintzberg (1979) has argued that the SimpleStructure is best characterized by what it is not.

290

CratraUsatlonCentralized Decentralized

Formalized

Fonnaliscrtion

Informal/ Simple JI Structure I

1 Machine \I Bureaucracy /

1 Professional Ay Bureaucracy y

Simple Complex Simple

CompUxlty

Complex

Figure 1. Relative dimensions of three structural types.

Specifically, it has little or no technical or ad-ministrative support staff, little differentiationbetween units, a "loose" division of labor, and avery small managerial hierarchy. In addition,"little of its behavior is formalized, and it makesminimal use of planning, training, and. . .liaisondevices" (p. 306). The Simple Structure is a formwhere all important decisions are centralized inthe hands of a dominant executive (CEO), whoinformally coordinates the organization's func-tional units. It is clearly a structure that is high incentralization and low in both formalization andcomplexity. Therefore, it is similar to Pugh et al.'s(1969) "implicitly structured organizations."

Since the Simple Structure is dominated bycentralization, its pattern of strategic processcharacteristics can be predicted from the propo-sitions provided earlier in the first column ofTable 2. For example, because power and knowl-

edge are concentrated, only the CEO can initi-ate a response to problems or opportunities. Thedominance of the CEO is also reflected in theexplanation of strategic action—organizationalactions reflect his or her intendedly rationalchoices. Similarly, with such a structure theactions that are chosen may be motivated by theproactive, personal interests of the CEO, andthey will be made to achieve a "positive," thoughgeneral, goal. This argument illustrates Mintz-berg's (1979) observation that a potential benefitof the simple structure is its "sense of mission."However, because the goal reflects the CEO'sF)ersonal preferences, the organization may con-tinue to pursue it long after it should have beenabandoned.

An organization with a Simple Structure is notrestricted by formalized procedures or forced tobargain among members who have different

291

preferences, so when it takes strategic actionsthey are more likely to be major departures fromits existing strategy. Moreover, the success orfailure of those actions can be directly attributedto the CEO because his or her cognitive limita-tions are the primary constraint on the compre-hensiveness of the strategic decision process. Inaddition, the CEO dominates the entire decisionprocess, and may well have an intimate knowl-edge of the firm's daily operations. Therefore, itis more likely that decisions will be consistentand integrated.

The Simple Structure is most successful in anenvironment that is, using Duncan's (1972) di-mensions, both simple (i.e., it has relatively fewcritical variables) and dynamic (i.e., those vari-ables are shifting). Its pattern of strategic pro-cess characteristics helps explain why. First, theCEO's dominance makes the organization di-rectly dependent on his or her preferences andcognitive capabilities, but the CEO has a realis-tic chance to understand a simple environment.Moreover, the high level of centralization equipsthe CEO with an understanding of both operat-ing and strategic-level issues, which when com-bined with the need for only one individual todecide, enables the organization to move quicklywhen faced with environmental change. There-fore, strategic decisions in a Simple Structuretend to be made quickly, in pursuit of positiveopportunities, and with a sense of direction andintegration. However, a complete dependenceon the CEO is a constant source of risk with thistype of structure.

Formalization and Machine Bureaucracy

A Machine Bureaucracy is a structure thatrelies on the standardization of work, whichmakes it similar to the structures previouslydescribed by Inkson, Pugh, and Hickson (1970)and Pugh et al. (1969). Its most distinguishingfeatures include ". . . very formalized proceduresin the operating core, a proliferation of rules,regulations and formalized communicationthroughout. . ." (Mintzberg, 1979, p. 315). In ad-dition, this structure tends to have large, func-

tionally grouped units at its lower levels, as wellas an elaborate administrative staff. Althoughthe work of those at the lowest levels may be themost directly controlled by formalization, it isimportant to recognize that ". . .at every hierar-chical level, behavior in a Machine Bureaucracyis relatively more formalized than in other struc-tural configurations" (Mintzberg, 1979, p. 318).

The propositions previously presented in thesecond column of Table 2 capture the pattern ofstrategic process characteristics that a MachineBureaucracy is likely to exhibit. For example,because this structure is above all else formal-ized, it is likely that the process will be initiatedonly when the condition of some formally moni-tored variable indicates a need for action. Inaddition, the strategic action ultimately taken willreflect the application of one standardized re-sponse from among those that the organizationhas developed. In combination, the above obser-vations also suggest that the action will be takento achieve a precise goal (e. g., a specified growthor profitability level), but that the goal will beremedial (i.e., a correction to the initially moni-tored deviation). However, because membersrecognize that their decision-making behavior issupposed to conform to specified rules and pro-cedures, there is an increased likelihood thatmeans will displace ends in a Machine Bureau-cracy.

As pointed out above, the dominating influ-ence of formalization in this structure explainsstrategic action as the output of standardizedorganizational processes. These institutionalizedprocesses have the added effect of producingstrategic actions that are only incremental depar-tures from the existing state. Moreover, the appro-priateness of the actions taken in a MachineBureaucracy is constrained, not by the CEO'scognitive limits, but by the level of detail that isachieved in its many systems (e.g., planning,information) and processes. However, becausesuch systems rely almost entirely on aggregated,quantitative data that must be passed throughmultiple levels, they can be expected to yieldonly a moderately integrated strategy.

292

A Machine Bureaucratic structure has its great-est success in a simple, stable environment; thecharacteristics of its strategic decision processsuggest why. An organization using Ihis struc-ture responds to formally monitored variablesand applies established standards in perform-ing its work. Therefore, the environment must besimple enough to allow critical variables to beidentified, and stable enough so that they canbe tracked and standards developed. However,if the environment changes, even the highestlevels of the organization may be unresponsivebecause the need for change may be masked bya dependence on information systems that gatherdata on a restricted range of variables. Moreover,there is an expectation among organizationalmembers that strategic decision making shouldproceed through formally established channels,which may contribute to costly delays in a chang-ing environment.

Complexity and Professional Bureaucracy

The "Professional Bureaucracy" is the namethat Mintzberg (1979) gave to the structure mostfrequently used in organizations such as gen-eral hospitals, universities, school systems, andsocial service agencies. These organizations relyon highly trained professionals who control theirown work, so the structure can accurately bedescribed as very decentralized. Similarly, be-cause the work requires detailed knowledge ofspecialized topic areas, the resulting structure ishorizontally complex and differentiated; verticaldifferentiation is limited. It should be emphasizedthat Professional Bureaucracies require stan-dardized behavior from their members, but thatbehavior is achieved much differently than in aMachine Bureaucracy. Members in this thirdstructural type are expected to enfer with skillsand behavior standards established by theirprofessions. However, these standards are notjust another type of formalization; they reflect aseparate, independent dimension—complexity/specialization (Reimann, 1973). (The terms com-plexity and specialization are often used inter-changeably in the literature, and they have been

shown by Grinyer and Yasai-Ardekani, 1980, tobe related.)

Since complexity is the dominant dimension ofa Professional Bureaucracy, the propositions pre-sented earlier in the final column in Table 2 high-light its pattern of strategic process characteris-tics. Specifically, strategic problems or opF>ortu-nities may go unrecognized or ignored becausemembers' interests are highV specialized, andtheir perceptions parochial. This high level ofhorizontal specialization also increases the likeli-hood that strategic action will be taken only afterextensive political bargaining among members,or as the result of individual members applyingsolutions from their collective "garbage can" ofskills (Cohen, March, &Olsen, 1972). In addition,diversity among members and the salience oftheir personal goals is likely to decrease theimpact of organization-level goals and producestrategic actions that are only incremental depar-tures from the organization's current state. Simi-larly, parochial perceptions are the primary con-straint on the comprehensiveness of the strate-gic decision process, and they contribute to mak-ing the integration of decisions quite low. Theseobservations are reflected in Mintzberg's (1979)conclusion that ". . .the notion of a strategy—asingle, integrated pattern of decisions commonto the entire organization—loses a good deal ofits meaning in a Professional Bureaucracy" (p.363).

The work performed by Professional Bureau-cracies is typically difficult to learn, yet quite welldefined (e.g., even complex surgical proceduresuse widely agreed-upon techniques). Therefore,the environment is accurately described as com-plex and stable. It is complex because it requiresskills learned only through advanced training,and stable because the necessary skills areenduring enough to allow the profession todevelop performance standards. In addition, thestrength and divergence of members' goalsmake such organizations highly political. There-fore, the only apparent way that the executive-level management of a Professional Bureaucracycan develop an overall strategy is by "patching"

293

together the disparate project and program pref-erences of the professionals, or by allocatingresources only to those that offer apparent syn-ergies.

Concluding ObservationsThis paper takes only a "first step" to encour-

age investigation and debate on the strategicprocess/structure question. It is suggested,though, that many critical issues have beenraised, and some answers have been offered.For example, firms have been observed whichexhibit consistent "patterns of behavior" in mak-ing strategic decisions (Fredrickson'& Mitchell,1984). The arguments presented here suggest thatstructure's pervasive impact offers a reasonableexplanation of why a firm develops a particularway of making strategic decisions. More impor-tantly, these same arguments also suggest thatalternative models of strategic decision making(Allison, 1971; Mintzberg, 1973) are more thanjust different perspectives on the same phenom-enon. Organizations that differ in their dominantstructure are likely to make strategic decisionsusing a very different process.

The arguments presented here also provide adescription that is richer than those previouslyavailable to explain why different structures havebeen associated with varying levels of perfor-mance in different contexts (Bums & Stalker, 1961 ;Lawrence & Lorsch, 1967; Khandwalla, 1977;Woodward, 1965). For example, it is the combin-ation of a dominant decision maker, who is pur-suing positive goals, is willing to make majordepartures, has detailed knowledge of the entireorganization, and is faced with a situation thatcan be understood, that enables some smallorganizations to succeed in a rapidly changingindustry in spite of being constrained by the deci-sion maker's cognitive limitations. This examplealso sheds light on the recurring "unit of analysis"issue that was most recently summarized byPfeffer (1982). More specifically, a firm that ishighly centralized is likely to have a strategicdecision process that is best understood by usingan individual unit of analysis, while an organiza-

tional perspective sheds light on the same pro-cess in a firm that is dominated by formalization.In contrast, the small group, with all its socio-political phenomena, is the basic unit of analy-sis for understanding the strategic process in anorganization whose dominant dimension is com-plexity.

This paper also encourages a broader debateby assuming an uncommon perspective on or-ganizational structure. Most strategic manage-ment scholars continue to see structure as sim-ply a tool for implementing strategy, while orga-nization theorists discuss the relative effects thatenvironment, technology, or size have on struc-ture. It is suggested that each of these views isunbalanced in its portrayal of structure. Thearguments presented here emphasize that a bal-anced view of the strategy/structure relationshipmust acknowledge that the strategic decision pro-cess and its outcomes can be facilitated, con-strained, or simply shaped by structure's directeffects. In accepting this argument, investigatorsare not being asked to reject evidence that led tothe "structure follows strategy" proposition. Theyare asked to recognize that there is a sizablebody of contributions which argues that anorganization's structure may have importantdeterministic effects of its own. These effects haveneither been widely recognized nor investigatedbecause the literature has been extremely frag-mented and underdevelopjed.

While this paper may have raised some criti-cal issues and offers a few tentative answers, itis only a first step. For example, the propositionspresented here describe the strategic decisionprocess in organizations where one of three struc-tural dimensions dominates. There may be cir-cumstances where these dimensions interact toproduce a strategic process whose pattern ofcharacteristics was not described. The empiricalquestions: when does structure follow strategy,and when does structure, through its direct effecton the strategic decision process, determinestrategy, remain. It is suggested that structure ismost likely to dominate in organizations wherean overall strategy is not institutionalized (i.e..

294

either no explicit strategy has been articulatedor it is in the process of changing). Similarly,there may be contexts such as crisis situationswhere the effects of structure are subservient tovariables such as environment. Therefore, it ishoped that investigators will not only challenge

and test the primary arguments and ideas pre-sented here, but that they will also refine andextend them. Such efforts are critical to under-standing the effects of structure and to develop-ing a balanced view of the strategy/structurerelationship.

References

Allison, G. T. (1971) Essence of decision: Explaining the Cubanmissile crisis. Boston: Little, Brown.

Blackburn, R. S. (1982) Dimensions of structure: A reviewand reappraisal. Academy of Management Review, 7,59-66.

Bobbitt, H. R., & Ford, J. D. (1980) Decision maker choice as adeterminant of organization structure. Academy of Man-agement Review. 5, 13-24.

Bourgeois, L. J., & Astley, W. G. (1979) A strategic model oforganizational conduct and performance. InternationalStudies of Management and Organization. 6(3), 40—66.

Bower, J. L. (1970) Managing the resource allocation pro-cess. Cambridge, MA: Harvard University Press,

Braybrooke, D., & Lindblom, C. E. (1970) A strategy ofdecision: Policy evaluation as a social process. New York:Free Press.

Bruner, J. S., Goodenough, J. J., & Austin, G. A. (1956) Astudy of thinking. New York: Wiley.

Burgelman, R. A. (1983) A model of the interaction of strate-gic behavior, corporate context, and the concept of stra-tegy. Academy of Management Review. 8. 61-70.

Bums, T., & Stalker, G. M. (1961) The management of innova-tion. London: Tavistock.

Carter, E. E. (1971) The behavioral theory of the firm and toplevel corporate decisions. Administrative Science Quar-terly. 16, 413-428.

Chandler, A. D. (1962) Strategy and structure: Chapters inthe history of the American industrial enterprise. Cam-bridge: MIT Press.

Child, J. (1972) Organizational structure, environment, andperformance: The role of strategic choice. Sociology. 6,1-22.

Child, J. (1974) Comments on Reimann and Mansfield's"bureaucracy." Administrative Science Quarterly. 19,247-250.

Cohen, M. D., March, J. G., & Olson, 1. P. (1972) A garbagecan model of organizational choice. Administrative Sci-ence Quarterly. 17, 1-25.

Crozier, M. (1964) The bureaucratic phenomenon. Englishtranslation. University of Chicago Press.

Cyert, R. M., & March, J. G. (1963) A behaviorai theory of thefirm. Englewood Cliifs, NJ: Prentice-Hall.

Delbecq, A., Van de Ven, A. R, &GustafeonD. H. (1975) Grouptechniques for program planning: A guide to NominalGroup and Delphi processes. New York: Scott, Foresman.

Duncan, R. (1972) Characteristics of organizational environ-ments and perceived environmental uncertainty. Ad-ministrative Science Quarterly. 17, 313-327.

Duncan, R. (1979) What is the right structure? Decision treeanalysis provides the answer. Organizational Dynamics.7. 59-80.

Fahey, L. (1981) On strategic management decision pro-cesses. Strategic Management Journal. 2. 43-60.

Fouraker, L. E., & Stopford, J. M. (1968) Organizational struc-ture and the multi-national strategy. Administrative Sci-ence Quarterly. 13, 47-64.

Ford, J. D., & Slocum, J. W. (1977) Environment, technologyand the structure of organizations. Academy of Manage-ment Review, 2. 561-575.

Fredrickson, J. W. (1983) Strategic process research: Oues-tions and recommendations. Academy of ManagementReview. 8, 565-575.

Fredrickson,}. W., & Mitchell, T. R. (1984) Strategic decisionprocesses: Comprehensiveness and performance in anindustry with an unstable environment. Academy of Man-agement Journal. 27, 399-423.

Fry, L. W. (1982) Technology-structure research: Three criti-cal issues. Academy of Management Journal. 25, 532-551.

Fry, L. W., & Slocum, J. W. (1984)Technology, structure, andworkgroup effectiveness: A test of a contingency model.Academy o/Management/ournaJ, 27, 221-246.

Galbraith, J. R. (1973) Designing complex organizations.Reading, MA: Addison-Wesley.

Gerwin, D., & Tuggle, F. D. (1978) Modeling organizationaldecisions using the human problem solving paradigm.Academy of Management Review. 3, 762-773.

295

Grinyer, P. H., & Yascri-Ardekani, M. (1980) Dimensions oforganizational structure: A critical replication. Academyof Management Journal, 23, 405—421,

Hage, J., & Aiken, M. (1967) Program change and organiza-tional properties. American Journal of Sociology, 72,503-518.

Hage, J., & Aiken, M. (1969) Routing technology, social struc-ture and organizational goals. Administrative ScienceQuarterly, 14, 368-379.

Hall, R. H. (1977) OrganizaWons.'Sfrucfure and process. Engle-wood Cliffs, NJ: Prentice-Hall.

Hedberg, B. L, T., Nystrom, P. C , & Starbuck, W. I. (1976)Camping on seesaws: Prescriptions for a self-designingorgdnization. Administrative Science Quarterly, 21,41-65.

Inkson, I. H. K., Pugh, D. S., & Hickson, D. I. (1970) Organi-zation, context and structure: An abbreviated replication.Administrative Science Quarterly, 15, 318-329.

Jelinek, N4^1977) Technology, organizations and contingencytheory. Academy of Management Review, 2, 17-26.

Khandwalla, P. N. (1977) The design of organizations. Har-court Brace Jovanovich.

Latham, G. P., & Yukl, G. A. (1975) A review of research onthe application of goal setting in organizations. Academyof Management Journal, 18, 824-845.

Lawrence, P. R., & Lorsch, J. (1967) Organization and envi-ronment. Boston: Harvard Business School.

Lenz, R. T., & Lyles, M. A. (1983) Crippling effects of "hyper-rational" planning. College of Commerce and BusinessAdministration, University of Illinois at Urbana-Cham-paign. Faculty Working Paper No. 956.

Lindblom, C. E. (1959) The science of muddling through.Public Administration Review, 19, 120-128.

March, J. G., & Simon, H. A. (1958) Organizations. New York:Wiley.

Merton, R. K. (1940) Bureaucratic structure and personality.Social Forces, 18(4), 560-568.

Miller, D., & Friesen, D. H. (1977) Strategy-making in con-text: Ten empirical archetypes. Journal of ManagementStudies, 14, 253-280.

Mintzberg, H. (1973) Strategy making in three modes. Cali-fornia Management Review, 16(2), 44-53.

Mintzberg, H. (1978) Patterns in strategy formation. Manage-ment Science, 24, 934-948.

Mintzberg, H. (1979) The sfrucfuring of organizations. Engle-wood Cliffs, NI: Prentice-Hall.

Mintzberg, H., Rcdsinghani, D., &Theoret, A. (1976)The struc-ture of unstructured decision processes. Administrative Sci-ence Quarterly, 21, 246-275.

Ouchi, W. G. (1978) The transmission of control through orga-nization hierarchy. Academy of Management/oumaJ, 21,173-192.

Perrow, C. (1972) Complex organizations: A critical essay.Glen view, IL: Scott, Foresman.

Pettigrew, A. (1973) The politics of organizational decision-making. London: Tavistock.

Pfeffer, J. (1982) Organizations and organization theory.Boston: Pitman.

Pugh, D. S., Hickson, D. J., & Hinings, C. R. (1969) An empiri-cal taxonomy of structures of work organizations. Ad-ministrative Science Quarterly, 14, 115-126.

Pugh, D. S., Hickson, D. J., Hinings, C. R., & Turner, C.(1968) Dimensions of organization structure. AdministrativeScience Quarterly, 13, 65-91.

Ouinn, J. B. (1978) Strategic change: Logical incrementalism.Sloan Management Review, 20(1), 7-21.

Ouinn, J. B. (1980) Strategies for change: Logical incremen-talism. Homewood, IL: Irwin.

Reimann, B. C. (1973) On the dimensions of bureaucraticstructure: An empirical reappraisal. Administrative Sci-ence Quarterly, 18, 462-476.

Rumelt, R. P. (1974) Strategy, structure and economic perfor-mance. Cambridge, MA: Harvard University Press.

Schendel, D. E., & Hofer, C. W. (Eds.). (1979) Strategic man-agement. Boston: Little, Brown.

Schwenk, C. R. (1984) Cognitive simplification processes instrategic decision-making. Strategic Management Journal.5, 111-128.

Simon, H. A. (1964) On the concept of organizational goal.Administrative Science Ouarteriy, 9, 1 -22.

Simon, H. A. (1974) How big is a chunk? Sdence, 183(4124),482-488.

Simon, H. A. (1976) Administrative behavior (3rd ed.). NewYork: Free Press.

Steinbruner, J. D. (1974) The cybernetic theory of decision.Princeton, NJ: Princeton University Press.

Thompson, J. D. (1967) Organizations in action. New York:McGraw-HUl.

Van de Ven, A. H. (1976) A framework for organizationassessment. Academy of Management Review, 1, 64-78.

296

Walton, E. J. (1981) The comparison of measures of organiza- complexity on implementation attempts. Administrative Sci-tion structure. Academy of Management Journal, 6, 155- ence Quarterly, 25, 428-440.160.

White, S., Dittrich, J. E., & Lang, J. R. (1980) The effects of Woodward, I. (1965) Industrial organizations: Theory andgroup decision making process and problem situation- practice. Oxford University Press.

James W. Fredrickson is Associate Professor of Business inthe Graduate School of Business, Columbia University.

297