1ceo Cognitive Map

8/6/2019 1ceo Cognitive Map

1/22

CEOs' Cognitive Maps and the Scope of the OrganizationAuthor(s): Roland Calori, Gerry Johnson, Philippe SarninSource: Strategic Management Journal, Vol. 15, No. 6 (Jul., 1994), pp. 437-457Published by: John Wiley & SonsStable URL: http://www.jstor.org/stable/2486760 .Accessed: 02/03/2011 20:22

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unlessyou have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you

may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at .http://www.jstor.org/action/showPublisher?publisherCode=jwiley. .

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed

page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of

content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms

of scholarship. For more information about JSTOR, please contact [email protected].

John Wiley & Sons is collaborating with JSTOR to digitize, preserve and extend access to Strategic

Management Journal.

http://www.jstor.org
http://www.jstor.org/action/showPublisher?publisherCode=jwileyhttp://www.jstor.org/stable/2486760?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/action/showPublisher?publisherCode=jwileyhttp://www.jstor.org/action/showPublisher?publisherCode=jwileyhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/2486760?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=jwiley


2/22

Strategic Management Journal, Vol. 15, 437-457 (1994)

CEOs' COGNITIVEMAPS AND THESCOPE OF THEORGANIZATIONROLANDCALORIGroupe ESC Lyon, Ecully Cedex, FranceGERRY OHNSONCranfield School of Management, Cranfield Institute of Technology, Cranfield,Bedford, U.K.PHILIPPEARNINInstitutde Recherche de l'Entreprise, Groupe ESC Lyon,Ecully Cedex, France

In this paper, CEOs are considered as 'cognizers' charged with integrating views in the topmanagement team; a role which should require high cognitive complexity especially indiversified multinational corporations. A methodology for studying top managers' cognitivecomplexity is described and then applied to a sample of 26 CEOs. The CEOs' cognitivemaps of the structure and of the dynamics of their industry are analyzed in terms of theirdegree of complexity, in relation to the breadth of the business portfolio of the firm, itsgeographic scope and the links the firm has with foreign parents. The results of thisexploratory test generally confirm the principle of requisite cognitive complexity, and reveala new set of more precise hypotheses linking particular dimensions of the scope of thefirmwith particular dimensions of CEOs' cognitive complexity.

Strategic problems and strategic decision makingare complex (Mintzberg, Raisinghani, and Theo-ret, 1976; Lyles and Mitroff, 1981; Mason andMitroff, 1981; Cray et al., 1991). Environmentsbecome more complex as international compe-tition and knowledge flows develop (Porter,1986; Melin, 1992); and the scope of somediversified multinational companies (DMNCs)adds even greater complexity (Prahalad andDoz, 1987). Firms, industries and businessenvironments may be considered as systems andtheir level of complexity can be conceived of interms of systems theory: the number of elementsin the system and the number of links between

Key words: Cognitive complexity, cognitive maps,CEOCCC0143-2095/94/060437-21(?)1994by JohnWiley & Sons, Ltd.

elements, representing the variety of the system(Ashby, 1956, 1958; Beer, 1966). The diversityof businesses in diversified companies (Prahaladand Bettis, 1986) and the diversity of countriesin multinational companies (Prahalad and Doz,1987) create high levels of variety, that topmanagers have to deal with in making strategicdecisions. In diversified companies there may besome synergies or sharing of resources but varietydominates: each business requires a specificstrategy corresponding to specific 'rules of thegame' played with specific actors. In international/multinational companies, the diversity of coun-tries (in terms of customer behavior and manage-ment practices) requires local adaptations; theremay be some global integration, but globalstrategy needs to be balanced with local demands(Prahalad and Doz, 1987).

Received 17 February 1992Final revision received 29 November 1993


3/22

438 R. Calori, G. Johnson and P. SarninMANAGING COMPLEXITYAccording to Chandler (1962) and Bower (1972)the complexity of diversified firms is resolved bymeans of organizational structures (such asdivisionalization), and by adequate administrativemechanisms. Businesses with similar strategiccharacteristics are grouped into 'sectors' or'strategic business units.' SBU managers makestrategic decisions for their business. The corpor-ate level influences the business unit level byorchestrating the organizational context. Argu-ably, the same kind of organizational design canbe applied to multinational activities (divisionscorresponding to geographic areas) in order toreduce the strategic variety that top managershave to deal with at the corporate level.However, according to Prahalad and Bettis(1986: 496) 'Organizational structure can attenu-ate the intensity of strategic variety that corporatelevel management must deal with, but it cannotsubstitute for the need to handle strategic varietyat the corporate level.' Prahalad and Doz (1987)also argue that while adequate organizationalstructures are necessary, they are not sufficientto manage diversified multinational corporations(DMNCs). In DMNCs, top managers have tocope with complexity through managing a globalmatrix, with a variety of management tools, butalso by developing a world view or mindset(Prahalad and Doz, 1987).Prahalad and Bettis (1986: 485) coined theterms 'dominant logic' to define the '. . . mentalmaps developed through experience in the corebusiness and sometimes applied inappropriatelyinother businesses.' When businesses are dissimilar,top management should presumably create thecapacity for multiple dominant logics to coexistand develop a 'meta-learning in which thedominant coalition learns to simultaneously con-ceptualize different type businesses' ((Prahaladand Bettis, 1986: 485). Ginsberg (1990) alsoargues that the management of diversified compa-nies requires an adequate level of 'socio-cognitive'complexity at the level of the top managementteam. According to Bartlett and Ghoshal (1989)managing a 'transnational' company has more todo with developing managers than with designingstructures and procedures: 'Diverse roles anddispersed operations must be held together by amanagement mindset that understands the needfor multiple strategic capabilities, views problems

and opportunities from both local and globalperspectives, and is willing to interact with othersopenly and flexibly. The task is not to build asophisticated structure, but to create a matrix inthe minds of managers' (Bartlett and Ghoshal,1989: 212).'Mental maps,' 'meta-learning,' 'mindsets;' allthese terms refer to the concepts of cognitivestructure and cognitive complexity of the topmanagers. In line with the arguments summarizedabove, we suggest that, as well as adoptingadequate administrative mechanisms, handlingdiversity requires high cognitive complexity fromtop managers in order to embrace the complexityof his or her environment. We focus on theparticular case and role of the chief executive(CEO), in this respect, we propose a methodologyto measure cognitive complexity, and apply it inan exploratory test of the relationships betweenthe scope of the company and the cognitivecomplexity of the CEO.

THE CEO AS A 'COGNIZER'The Chief Executive has been characterized asa decision-maker (e.g., Learned, Christensen,and Andrews, 1961), as a visionary leader (e.g,Bennis and Nanus, 1985), and as a political actor(e.g., Greiner, 1986). In their claim for moreresearch on 'upper echelons' Hambrick andMason (1984) argue that the strategies and theeffectiveness of an organization are reflectionsof the values and cognitive bases of its powerfulactors, among whom is the strategic leader (seealso Schwenk, 1988). Strategies are abstractionsin the mind of managers (Mintzberg, 1987); theyemerge from sets of ideas and constructs throughwhich problems are identified and interpreted(Hedberg and Jonsson, 1977). Although severalactors may interact through social interchange toproduce collective interpretations, top manage-ment is responsible for providing organizations'interpretations of their environment and strategicresponses (Daft and Weick, 1984; Smircich andStubbart, 1985). 'Indeed, the imposition ofmeaning on issues characterized by ambiguityhas become a hallmark of the modern topmanager' (Thomas, Clark, and Gioia, 1993: 240).

Decision-makers construct simplified mentalmodels when dealing with complex problems(Simon, 1957; March and Simon, 1958; Simon,


4/22

CEOs' Cognitive Maps 4391976). They may be subject to selective perceptionsince they are unable to evaluate comprehensivelyall variables relevant to a decision (Tversky andKahneman, 1974; Hogarth, 1980; Mason andMitroff, 1981; Schwenk, 1988). For instance,research has identified simplifications and biasesin executives' maps of their industries (Shrivastavaand Lim, 1984; Stubbart and Ramaprasad, 1990).Such simplifications and heuristics are usefulhowever. Without such cognitive structuresdecision-makers would become paralyzed by theneed to analyze extensive ambiguous data (Weick,1979; Daft and Weick, 1984; Hogarth, 1980;Walsh and Fahey, 1986). The simplified abstractcognitive structures developed by managers helpthem to cope with decision-making complexities(Kiesler and Sproull, 1982).This duality raises the issue of the adequatelevel of cognitive simplicity vs. complexity ina given situation, and the question of thedevelopment of individuals' cognitive structures.Confronted with new or changing environments,managers are likely to deal with the situationwithin the bounds of existing cognitive structures(Johnson, 1988). It is typically after a phase ofunlearning (Hedberg, 1981), perhaps promptedby threat or crises, that managers reconceive ofsituations outside that cognitive structure (Argyrisand Schon, 1978, Grinyer and Spender, 1979).Some environmental events will be retained andincluded in the knowledge structure thus causinga new association to be developed (Meyer, 1982).Simple structures may therefore develop overtime into complex ones by the inclusion of newexperiences (Schank and Abelson, 1977; Lylesand Schwenk, 1992).As noted by Walsh and Fahey (1986) there isno agreed terminology to denote the key constructof this theory of managerial cognition: mentalmaps, frames of reference, mindsets, cognitivebase, beliefs, schemata, cognitive structures,cognitive maps are all mutually substitutablelabels. In line with Huff (1990) we will use theterm 'cognitive map' except when referring to aparticular author, in which case the original termused by the author will be preferred. There ismore agreement in defining individuals' cognitivecomplexity, according to complexity theory, thecomplexity of an individual conceptual system isdetermined by two interdependent aspects: thenumber of parts or dimensions of the system andthe nature and extent of rules for integrating

these parts (Schroder, Driver, and Streufert,1967; Streufert, 1972, 1973; Stabell, 1978). Theability to perceive several dimensions in astimulus array is referred to as differentiationor comprehensiveness and the development ofconnections among the differentiated character-istics is referred to as integration or connectedness(Bartunek, Gordon, and Weathersby, 1983).

REQUISITE COGNITIVE COMPLEXITYStudying the relationship between the cognitivecomplexity of the chief executive and managerialeffectiveness or firm performance raises sometheoretical issues. Decisions are often taken bya top management team, going through a politicalprocess of bargaining among the members of theteam and other stakeholders (Ford and Baucus,1987). In consequence it may be more appropriateto study the socio-cognitive complexity of theteam (Ginsberg, 1990), and the processes ofinfluences within the team in trying to reachconsensus. As Prahalad and Bettis (1986) argue:the variety of dominant logics that a topmanagement team (TMT) can handle dependson the composition of the team.

However, we suggest that the CEO has aparticular integrative function within the topmanagement team. Whatever the level of agree-ment and the distribution of power in the TMT(Walsh and Fahey, 1986) the chief executive hasto integrate views. High cognitive complexityand diversity among the team can be highlydysfunctional unless integration is achieved(Crossan, 1991). Particularly socio-cognitive com-plexity may be dysfunctional in the implemen-tation of strategy (Ginsberg, 1990). Given his/her position in the TMT, the CEO can, then,be viewed as a cognitive integrator, the architectof a 'congregate map which must match thevariety, equivocality and crypticality of thecombined strategy and social system phenomena,'(Bougon, 1992: 385). In order to play thisintegrative role the CEO may need high cognitivecomplexity as suggested by the argumentsdeveloped by Bartunek et al. (1983: 274). Peoplewith a higher level of cognitive complexity tendto have a higher tolerance for ambiguity, assumeleadership roles (Streufert, Streufert, and Cas-tore, 1968), are more capable of taking theperspective of others (Triandis, 1977), and of


5/22

440 R. Calori, G. Johnson and P. Sarninresolving conflicts cooperatively (Eiseman, 1978).In short, the tolerance for divergent perspectivesmay be crucial for CEOs in TMTs facing complexenvironments; and the leader's level of cognitivecomplexity should have significant effectsthroughout the organization (Bartunek et al.,1983).In his study among chief executives of 24 smallSwedish firms involved in the construction ofprefabricated houses, Hackner (1991) found thatthe relationship between the CEOs' cognitivecomplexity and the performance of the firmfollowed an inverted U-shaped curve. In thatresearch the degree of complexity of the environ-ment was kept constant across cases. However,the degree of environmental complexity mayinfluence the degree of cognitive complexity(Hedberg and Jonsson, 1978). Thus the relation-ship between cognitive complexity and perform-ance should be moderated by the degree ofenvironmental complexity. In line with Ashby'sLaw of Requisite Variety (1956; 1958) and withWeick (1979), we argue that, in order to beeffective, CEOs need to develop the ability togenerate several interpretations of events so thatthe variety in their understanding matches thevariety of the situation. In other words, incomplex situations high cognitive complexityshould lead to more accurate perception andmore effective behavior (Bartunek et al., 1983).In building and testing a theory of themoderating effects of environmental complexityon the relationship between CEOs' cognitivecomplexity and performance, the first step is toanalyze the relationship between the level ofenvironmental complexity and the CEOs' cogni-tive complexity. This is the aim of our research.The complexity of the environment of theCEO can be considered in two ways: the scopeof the company, defined here as the number ofbusinesses and the geographic scope, whichdetermines the number of competitive systemsin which the firm is involved; and the intrinsiccomplexity of each of these competitive systems.As it is difficult to measure the intrinsic com-plexity of a competitive system objectively, wemeasured complexity by an objective assessmentof the scope of the company.The relationship between the scope of thecompany and the cognitive complexity of theCEO should be twofold: a broad scope shouldrequire higher cognitive complexity, as suggested

by Prahalad and Bettis (1986) and Prahalad andDoz (1987); and a broad scope should be asource of learning from diversity, as suggestedby Hedberg and J6nsson (1978).In order to demonstrate the relationshipsbetween the scope of the organization and thestructure of top managers' cognitions, individualfactors which could affect executives' cognitivecomplexity should ideally be controlled. Individ-ual factors include: the executives' personalbackground (Dearborn and Simon, 1958; Martin,1982; Fiske and Taylor, 1984; Hambrick andMason, 1984; Haley and Stumpf, 1989; Hitt andTyler, 1991), their cognitive style (Hurst, Rush,and White, 1989) and their personal network ofsocial interactions (Schrank, 1978).

Both environmental circumstances and individ-ual characteristics shape managers' cognitivemaps. Whilst we recognize the importance of theindividual factors, this research focuses on thelink between the scope of the company (as ameasure of environmental complexity for theCEO) and the cognitive complexity of the CEOswho have to cope with variety.

MEASURING COGNITIVECOMPLEXITY: INTERVIEWS ANDCOGNITIVE MAPPINGThe first methodological choice was betweencontent free measures of cognitive complexity(see Schroder et al., 1967) and measures basedon the content of the issue addressed in theresearch-here the CEO's understanding of theenvironment. We opted for the second solutionconsidering that the interpretation of the relation-ships would be more instructive and that theinterest manifested by the CEOs would be higher.The CEO's understanding of his/her businessenvironment (or industry) can be analyzed intwo ways. Lenz and Engledow (1986) considerenvironmental analysis models in terms of'assumptions about environmental structure' and'origins/processes of environmental change.' Simi-larly, in the field of cognitive analysis, Piaget(1987) made a distinction between 'figurativestructures,' which are mental representations ofa structure and 'operative structures,' which aremental representations of the transformationprocess of a structure. We studied these twointerdependent aspects: the CEO's understanding


6/22

CEOs' Cognitive Maps 441of the structure of the environment of the firm(in terms of actors in the system and linksbetween actors), and the CEO's understandingof the dynamics of the environment (in terms offorces and relationships between forces).

Cognitive mapping has proved to be a usefulway of representing strategists' understanding ofenvironment and industryforces (Schwenk, 1988).Huff (1990: 41) suggests that mapping managers'interpretations of the environment can help toshow how 'elements of the environment arecategorized and evaluated' and 'indicate how theinitial definition of a situation changes overtime.' Cognitive mapping has been used forinstance to uncover competitive positioning asperceived by managers (Porac, Thomas, andEmme, 1987; Thomas and Porac, 1990; Reger,1990). According to Huff and Fletcher (1990:403-404): 'Cognitive mapping is a form of contentanalysis... (but) in cognitive mapping it is therelationship between cognitive elements that isbeing studied. Some cognitive maps are subjectto quantitative analysis, but the central benefitof the cognitive map, is that it encourages holisticsynthesis rather than reductive analysis of theactor's view of the world.'Two aspects of a cognitive map can be analyzed(Reger, 1990; Hackner, 1991):

the content of the map in the sense of themeaning it embodies,the configuration of the map, for instance interms of cognitive complexity (Streufert andSwezey, 1986).

In this research, we are only interested in theconfiguration of the maps. Indeed maps such asthose presented in Figures 1 and 2 give a directview of the level of complexity (number ofconcepts and number of links). For each managertwo maps could be drawn:

a map representing the perceived structureof the environment (cf. Figure 1),a map representing the perceived dynamicsof the environment (cf. Figure 2).We considered several approaches to elicitingcognitive maps. Prahalad and Bettis (1986) argue

that a person's theory in use cannot be obtainedsimply by asking for it and recommend the use ofcreative questionnaires and analysis procedures.

Ginsberg (1989: 417) notes 'the paucity ofmethods suitable for operationalizing the subjec-tive characteristics of managerial mental mapsinto quantitative and reproducible measures;'and proposes the use of repertory grid techniquesin order to elicit and quantify top managers'cognitive maps. Repertory grid techniques (Kelly,1955) have been used to elicit constructs andmap competitors according to these constructs(Reger, 1990). Repertory grid techniques aremore rigorous than in-depth open-ended inter-views, but some researchers report boredomfrom the interviewees (Brown, 1992; Reger andHuff, 1993). The 'Self-Q' Technique developedby Bougon (1983) could be useful in analyzingtop managers' understanding of the dynamics oftheir environment, if the number of constructswere not limited in the phase of elicitation ofthe relationships between constructs. Ladderingtechniques (Fransella and Bannister, 1977; Hin-kle, 1965) could also be used in interviewingmanagers on cause-effect relationships describingthe dynamics of their environment. But againthese are lengthy procedures. In order to coverthe scope of our research-perceptions of thestructure (competitors, markets, otherstakeholders) and of the dynamics of theenvironment-we would have had to combinethese techniques. Interviews following such proto-cols would have been excessively long. SomeCEOs may be reluctant to follow the protocols,or may shorten the interview. As acceptance bya pre-determined sample of chief executives wascrucial to the purpose of this research, we optedfor a more simple interview method.We chose in-depth open-ended interviewswhich are less rigorous, but are well accepted byCEOs and allow the elicitation of the mostsalient concepts. Such interviews are widely usedby researchers analyzing managers' cognitivemaps (among others: Hackner, 1991; Cossetteand Audet, 1992; Eden and Ackermann, 1992).In order to reduce the possible biases fromin-depth loosely structured interviews severalprecautions can be taken (see Eden and Acker-mann, 1992). These include: allowing sufficienttime for discussion; controlling the skills of theinterviewer; correcting biases introduced by theinterviewer; controlling the coding skills ofthe analysts; using coders with similar skills;establishing a precise protocol for coding; definingas precisely as possible the level of analysis;


7/22


8/22

CEOs' Cognitive Maps 443product terms and in publishing with regard tolanguage). Further, we sought an equal numberof firms in each industry in the two countriesand a split between large and small firms; andspecialist vs. generalist firms. The sample isshown in Table 1.Data on the scope of the company werecollected from documentary sources (annualreports, industry studies and the business press).The diversity of the business portfolio of afirm in the industry was evaluated by the numberof products x markets, in terms of threecategories: focus on one business/a few businesses/most of the businesses of the industry.The geographic scope of a firm was evaluatedby the geographical limits of its targeted marketsin the industry, according to three categories:national/national and a few foreign countries/international (several major foreign countries).

The interactions with a parent company wereevaluated according to three categories: indepen-dent company/belongs to a group from the samecountry/belongs to a multinational group from aforeign country.Top managers' perceptions of the environmentThe data on top managers' understanding of thebusiness environments were collected by openended interviews as explained earlier. Twenty-six chief executives participated (24 CEOs and 2General Managing Directors). The interviewstook place in the period from January 1990

Table 1. Sample of companiesRetail Book

Scope of the company Brewing Car banking publishing TotalFocus on one industry segment 1 1 3 5A few segments 3 1 1 5All segments in the industry 5 2 6 3 16National 3 3 2 8National and a few foreign countries 1 1 4 3 9International (several major foreign countries) 5 2 2 9Independent 3 4 4 11Belongs to a national group 3 1 3 1 8Belongs to a multinational group 3 2 2 7

9 3 7 7 26

to April 1990. The interviews were looselystructured, though grounded on two broadquestions designed to surface the CEOs' strategicthinking about their industry and their firm:

(a) 'What main changes do you expect inyour industry in the 90s?'(b) 'What changes are you thinking of foryour company in the 90s?'Though both questions were about change,managers also discussed the present configurationof the industry. In this way they surfaced

explanations of both structure and dynamics interms of this research. The interviews lastedfrom one hour and a half to two hours and ahalf, and all were tape-recorded and translatedinto English. In general about two-thirds of eachinterview were concerned with the first question,though many managers mixed discussion oftheir company with discussion of industry andcompetitive forces, confirming that the conven-tional distinction between environmental andorganizational factors is typically blurred (Dill,1962). Many of the interviews were highlydiscursive in nature with managers ranging overissues they chose to emphasize. Where managerswere less forthcoming two forms of promptswere employed: first, neutral conversationalprompts; and second more specific prompts basedon conventional framework of industry analysis:competitors, customers, suppliers, new entrants,substitute technologies, state(s) intervention(s)(Porter, 1980), and sources of competitive advan-


9/22

444 R. Calori, G. Johnson and P. Sarnintage and industry segmentation (Porter, 1985).However in order to reduce the biases createdby such specific prompts, we distinguished spon-taneous answers from prompted answers in theanalysis, deleting all answers to prompts whichdid not prove to be a crucial point further in thediscussion.Content analysisContent analysis was used to surface cognitiveconcepts and links beween such concepts (Holsti,1969). The analysis of the data involved foursteps.Phase one: Surfacing first order concepts andlinksInterview transcripts were analyzed according tothe four broad analytical dimensions defined exante: concepts describing the industry structure,links between concepts describing structure,concepts describing the industry dynamics andlinks between concepts describing the industrydynamics.As suggested by Stubbart and Ramaprasad(1990) the level of analysis (in terms of abstrac-tion, and aggregation) is a key issue in contentanalysis. We only retained 'core-constructs' whichwere central to the reasoning of the managers(Eden and Ackermann, 1992) and we gave anoverview of the cognitive map, by collapsingdetails into larger cognitive units (Anderson,1983). Links may sometimes just reflect thenature of verbal argument in interview conditions,so we only retained links which were explicitlymentioned by the interviewee. For each analyticaldimension first order concepts and links wereidentified ex post by two analysts workingseparately on each transcript. The first orderconcepts and links (or codes) were identified interms of the exact wording used by the informants(cf. Van Maanen, 1979).Discrimination in coding between 'structure'and 'dynamics' was undertaken according toprecise conventions.

all actors or groups of actors in the systemmentioned by the manager were consideredas elements of the structure of the industry;then explicit mention of links were soughtfor,

-all actors, groups of actors or other character-istics of the system which were explicitlyassociated with the notions of 'change,''transformation,' 'evolution,' 'development,''variation,' and their synonyms were con-sidered as elements of the dynamics ofthe industry; then explicit mention of linksbetween these elements were sought for.

In some cases the same concept could appearamong the elements of the perceived structureand the elements of the perceived dynamics bya given manager, if, and only if, it representedan actor explicitly involved in the transformationof the industry.The elements of Figures 1 and 2 are examplesof concepts and links which emerged from thecontent analysis of the interviews. For instance(cf. Figure 1) for the CEO of one of the retailbanks, the control by the U.K. Mergers andMonopolies Commission is one of the conceptsin the structure of the industry; and the alliancebetween the Royal Bank of Scotland and theBank of Santander is one of the links consideredas important in the structure of the industry. Forthe same CEO (cf. Figure 2) the vulnerability ofbuilding societies is one of the forces he seesdriving the dynamics of the industry and itspositive impact on new competitors' entry on theBritish market is one of the links in the dynamicsof the industry. Altogether the 26 interviewscontained 301 different first order concepts ofindustry structure and 210 different first orderconcepts of industry dynamics.Phase two: Weighting conceptsThe importance of each concept for each managerwas also assessed. Each concept was weightedaccording to four criteria: explicit mention of itsimportance by the manager; spontaneity; priorityin the interview; and relative length of thediscussion on the theme. This weighting is shownin Figures 1 and 2 by the use of asterisks, fourasterisks meaning high on the four criteria. Infurther analyses we retained the concepts whichscored on at least one of these criteria.Possible biases were reduced in two ways.First it was recognized that the interviewersthemselves could have introduced bias into thedata by prompting a particular theme during theinterviews. In order to reduce such bias, aconcept mentioned after a prompt was retained


10/22

CEOs' Cognitive Maps 445only if it scored in two of the three other criteria(explicit mention of importance by the manager,priority in the interview, and length of thediscussion on the theme). In order to reducebias in content analysis, the identification ofconcepts and weightings were undertaken sepa-rately by two analysts and then compared (butwe did not have the opportunity to appoint athird 'candid' analyst, i.e., unaware of thehypotheses). Where weightings differed by afactor of 2 or more, the weightings were averaged.However, differences by a factor of 2 or morewere only found for a total of 31 concepts across11 of the 26 cases; and differences such asconcepts identified vs. not identified were foundfor a total of 15 concepts across 6 of the 26cases. Considering these levels of divergenceintercoder reliability was 90.8 per cent.Phase three: Surfacing second order categoriesSecond order categories were identified ex postfrom the list of first order concepts and links.The second order categories represented attemptsto classify concepts at a higher level of abstraction.The following second order categories of conceptsdescribing the industry structure emerged: thecompany, customers, distributors, competitors,suppliers, regulators (government, EEC, etc.)and lobbies. Five second order categories of linksbetween concepts of industry structure werefound; dependence, direct competition, pressure,cooperation, similarities and differences inbehavior. Six second order categories of conceptsdescribing the industry dynamics were found:changes in context, strategic moves of actors,changes in structure, technology, industry outputand key success factors. Three second ordercategories emerged describing the links betweenconcepts of industry dynamics: proximity, equiva-lence of concepts (similarities or differences) andcausal relationships (positive or negative). Thesecond order categories were only used toorganize the cognitive maps in the fourth phase;the elements of the cognitive maps being thefirst order concepts and links.Phase four: Cognitive mappingTwo maps were drawn for each CEO. The firstmap on the perceived structure of the industrycorresponds to generic families 1 and 2 described

by Huff (1990): 'Maps that assess attention andimportance of concepts' (family 1) and 'Mapsthat show dimensions of categories and cognitivetaxonomies' (family 2). The second map on theperceived dynamics of the industry correspondsto the generic family 3 described by Huff:'Maps that show influence, causality and systemdynamics'. Figures 1 and 2 show two cognitivemaps from the 52 that were drawn. Figure 1shows the perceptions of the structure of theindustry from the CEO of a retail bank andFigure 2 shows the perceptions of the dynamicsof the industry from the same CEO.Each cognitive map was drawn using the firstorder concepts and links between conceptsdiscussed by the manager and organized accordingto the second order categories found in phasethree. For instance all the competitors (category)are positioned in a particular zone of the mapsof the structure, all the regulators (other category)are positioned in another particular zone of themap of the structure. The links between conceptsof the dynamics of the industry are drawnaccording to the following categories:- proximity ()- equivalence ( =-or # )- causality (- + or - -The links between actors in the system arerepresented by lines with a mention of the typeof relationship: (-O-)Measuring complexity: Comprehensiveness andconnectednessCognitive maps give a direct view of thecomplexity of the perception of a manager andallow straightforward calculations of scores ofcomprehensiveness and connectedness. Lang-field-Smith and Lewis (1989) give a review ofpossible measures.From the possible measures, we chose the bestsuited to a visual evaluation of maps.* The complexity of a CEO's cognitive maps ofthe structure of the environment was measured

by three variables:-the comprehensiveness of a manager's map

of the structure of the industry was measuredby the number of elements (concepts) in themap (cf. Langfield-Smith and Lewis, 1989),


11/22

446 R. Calori, G. Johnson and P. Sarnin

SHANGHAI ** AUSTRALIAN.A.B. ** JAPANESEHONG KONG Stakes ~ ~ ~ ~ ~ ~ ~ 'ak isigi T:k~

***BRITISHCOMPETTlORS G**GERMANMIDLANDBANK DEUTSCHEBANKROYAL BANK OF SCOTLAND (lots of hiddenreserves)BUILDINGSOCIETIES vulnerable) * ITALIANST.S.B. (vulnerable) Ml%e (Backwrd)

** SUPERMARKETS ** SPANISHCOMPETITORS* BANK OF SANTANDER(overbranched,profitable)** BANK OF ENGLAND(littleregulation) ** COMPANYRKI

(one of themajorEuropean ** FRENCH** MERGERSAND retail banks) (notprofitable,MONOPOLIES Stateorganizations)COMMISSION**EEC(definesmarketrules)***UK Market(peoplehave a strongtendency to borrow)

** JAPANESEMarket(more stocks,shares,savings, pensions)** SPANISHMarket(profitable)

** GREEKMarket(profitable)

Comprehensiveness tructure= 19Connectednessstructure = 5119= 0.26Complexitystrwture = 19 x 0.26= 5

Figure 1. Cognitive map/structure (Retail banking). CEO firm RK1

- the connectedness of a manager's map of thestructure of the industry was measured bythe number of links between concepts dividedby the total number of concepts in the map(cf. Hiackner, 1991),

- we also used a combined measure ofcomprehensiveness and connectedness-i.e., complexity-of a manager's cognitivemap of the structure of the industry: thenumber of links between concepts in the


12/22


13/22

448 R. Calori, G. Johnson and P. Sarninthe map (concepts) linked together directlyor indirectly; the largest submap is the onewhich includes the most elements linkedtogether,

we also used a combined measure of compre-hensiveness and connectedness-i.e.,complexity-of a manager's map of thedynamics of the industry: the number ofconcepts in the largest submap (i.e., the scoreof comprehensiveness multiplied by the scoreof connectedness).

For illustrative purposes Figures 1 and 2 alsogive the scores on the six variables for the CEOof a retail bank.CEOs' cognitive maps of the structure of theirindustry were found to include from 9 (minimum)to 34 (maximum) concepts, the average being17.9 and the standard deviation 6.6. The numberof links pointed out between actors varied from2 (minimum) to 7 (maximum), the average being3.5 and the standard deviation 1.1. The numberof links explicitly mentioned between concepts(here actors in the system) was relatively low.

This may, of course, be partly a consequence ofour interview and analytical conventions giventhat we did not prompt on possible links betweenactors, or count as links the perceived similaritiesbetween the elements of a category mentionedby the interviewers: for instance, in Figure 1,the similarities between the four companiesmentioned as 'British competitors.'

CEOs' cognitive maps of the dynamics of theirindustry included from 6 (minimum) to 16maximum concepts, the average being 9.9 andthe standard deviation 2.6. The average numberof concepts included in the managers' cognitivemaps of industry dynamics was similar to thenumbers found by Spender (1989) or Hiickner(1991). These numbers may seem low whencompared with some other cognitive maps (seeEden and Ackermann, 1992); however this maybe explained by the loose structure of theinterviews and most of all by our decision tocollapse details into core constructs for analyticalpurposes. Nonetheless, it seems that a limitednumber of core constructs is sufficient to explainthe dynamics of a competitive system, at leastfrom the point of view of a CEO. The largestsubmaps of the dynamics of an industry includedfrom 3 (minimum) to 16 (maximum) concepts;

the average was 7.5 and the standard deviation3.3 (i.e., on average 7.5 concepts of the mapwere linked together, either directly or indirectly,in the largest submap).In order to demonstrate the relationshipsbetween the scope of the organization and thestructureof top managers'cognitions, other possiblesources of cognitive complexity such as thecomplexity of the industry and individual factorsshould ideally be controlled. Here we were ableto control for the complexity of the industry, butnot for the managers' personal characteristics;because of the high diversity of the top managers'backgroundsand the limitednumberof respondentsin our study. We acknowledge that our inabilityto control for individual factors may limit theexplanatory power of our findings.Controlling for the complexity of the industryIndustries vary in their complexity; as the topmanagers interviewed belonged to companies infour different industries, industry complexity hadto be controlled for. For instance, in the carindustry, global competition, technological issuesand networks of alliances may increase thecomplexity of the competitive system; on theother hand the book publishing industry is stillmainly domestic and relatively protected bylinguistic and cultural barriers. So, in order tostudy the relationships between the scope of theorganization and the configuration of the topmanagers' cognitive maps, we standardized thevariables measuring comprehensiveness, connec-tedness and complexity. The standardization wasdone by dividing the original values of thevariables by the average for the correspondingindustry. For instance, the comprehensiveness ofa manager's cognitive map of the structure ofthe industry 'i' is:

number of concepts in the manager's cognitive mapaverage number of concepts in the cognitive maps of the managers in industry 'i'

And so on for all the variables measuringcomprehensiveness and connectedness. Stand-ardized measures were used in all furtheranalyses.Controlling for the length of the interviewIt was recognized that differences in the lengthof the interviews could bias the measurement


14/22

CEOs' Cognitive Maps 449of comprehensiveness: the longer the interview,the more concepts in the map. Several factorscould affect the length of the interview: onemanager might have more to say than another(though in this case there would be no bias);the manager's agenda might have curtailed orallowed a lengthier interview; and of coursethe behavior of the interviewer may influencethe length of the interview. The length of theinterview was measured by the number ofwords. Actually correlations between the lengthof the interview and measures of comprehen-siveness and connectedness only showed asignificant relationship with the comprehen-siveness of the maps of the dynamics(r = 0.48*). The probabilities of significantrelationships with the other dimensions ofcomplexity were low. However, we decided tocontrol for the length of the interview in allfurther analyses in order to produce morereliable results and to correct possible biasesfrom the respondent and from the interviewer.Testing relationships between dependentvariablesWe assumed that comprehensiveness and con-nectedness were not related to each other.However, in order to test for the possiblerelationships between comprehensiveness andconnectedness we computed standardized par-tial correlation coefficients between the twovariables after controlling for the length ofthe interview. The results of the multipleregressions show that the measures of compre-hensiveness and connectedness are not posi-tively correlated. The comprehensiveness andthe connectedness of the maps of the structureare weakly and negatively correlated(Beta = -0.43*); the comprehensiveness andthe connectedness of the maps of the dynamicsare not correlated (Beta = 0.02). On the otherhand (as expected), the combined measure ofcomplexity of the maps of dynamics (numberof concepts in the largest submap) correlateswith both comprehensiveness (Beta = 0.77***)and connectedness (Beta = 0.66***). Asexpected the combined measure of complexityof the maps of structure (number of linksbetween concepts) correlates with comprehen-siveness (Beta = 0.41*) and with connectedness(Beta = 0.62**).

Data analysisThe variables describing scope were all nominallevel; the variables measuring standardized com-prehensiveness, connectedness and complexitywere ratio level. As we had to control for thelength of the interviews, 18 analyses of covariancewere performed (ANCOVA) with (successively)each dimension of cognitive maps by(successively) each dimension of scope, with thelength of the interview as a covariate. Whensignificant differences were found, Duncan'smultiple range test was computed to comparegroups.

RESULTSThe results of the analyses of covariance isolateseveral significant relationships between theCEOs' cognitive complexity and the scope of theorganization; the presentation of the resultsfocuses on these, and on one relationshipapproaching significance. The relatively low levelsof variance explained probably arises from thesmall size of the sample and from the uncontrolledvariance due to individual factors and specificfirms' strategies.Cognitive complexity and diversity of theportfolio of businessesTable 2 summarizes the results concerning therelationships between the CEOs' maps of thedynamics of the industry and the diversity of theportfolio of businesses.Hypothesis 1-1 is not supported, the differencesin the average scores of comprehensivenessbetween the three categories (focused/severalbusinesses/all the businesses in the industry) arenot significant, however they seem to follow theexpected pattern with high comprehensiveness indiversified firms and low comprehensiveness infocused firms. The differences between theaverage scores of connectedness between thethree categories approach significance. In firmswith a diversity of businesses the CEOs' cognitivemaps of the dynamics of the industry seem tobe less connected than the maps of the CEOs infocused firms, in line with Hypothesis 1-2.The ratio measure of connectedness (numberof concepts in the largest submap divided by the


15/22

450 R. Calori, G. Johnson and P. SarninTable 2. Diversity of businesses and complexity of the maps of the dynamics of the industry

Standardized comprehensiveness/Dependent variable dynamics Standardized connectedness/dynamicsFactor Diversity of businesses in the industry Diversity of businesses in the industryCovariate Length of the interview Length of the interviewMain effect F = 0.70 sig: 0.51 F = 2.30 sig: 0.12Covariate F = 1.73 sig: 0.20 F = 0.00 sig: 0.96Multiple R2 R2 = 0.12 R2 = 0.17

nb of cases Mean S.D. nb of cases Mean S.D.Gi: focused firms 5 0.89 0.21 5 1.19 0.22G2: several businesses 5 1.00 0.17 5 0.97 0.09G3: all businesses in theindustry 16 1.03 0.27 16 0.95 0.24Duncan's multiple range Group 1 > Group 3test (under the 0.10 level)

total number of concepts) should be borne inmind. It may be that the diversity of products-markets leads to CEOs coping with such diversitythrough what amounts to chunking (Baddeley,1983), leading to a number of submaps. A visualcomparison of the maps in the two extremecategories (focused/diversified) confirms thisfinding. The existence of several submaps discon-nected from each other results in a low score ofoverall connectedness. In the terms used byPrahalad and Bettis (1986: 498) it shows that topmanagers in diversified companies 'simul-taneously conceptualize different type business-es,' they have in mind several coexisting logics.

Considering that the scores of comprehen-siveness were not significantly higher in diversifiedcompanies, the results suggest that chief execu-tives simplify their cognitive maps of the dynamicsof each business in diversified firms. Thisfinding also reveals a complementary measureof comprehensiveness (differentiation), whichshould be used in further research: namely thenumberof submaps (subsystems) in the manager'scognitive map.Cognitive complexity and geographic scopeIn firms with an international geographic scopethe CEOs' cognitive maps of the structure of theindustry are more comprehensive than those ofCEOs in firms with a national scope. This findingsupports Hypothesis 2-1. Table 3 gives moreinformation on this significant relationship.

Chief executives of firms which serve aninternational market have more competitors andmarket segments to consider and the diversity ofworld markets leads to more strategic groups ofcompetitors and market segments in the cognitivemaps. This result might seem self-evident; weargue it is not. As the opening question of theinterview was very broad, CEOs in internationalcompanies were free to simplify or reduce theirviews of the structure of the market as theywished. CEOs in companies with a nationalscope were free to widen their views of thestructure of the industry to the internationalscene: and it is quite conceivable they might doso bearing in mind that they could be competingin their home market with international competi-tors or, more generally, that they may be affectedby international actors (for instance the EEC,or international distributors) in a world economy.This result also shows that the experience ofinternational operations could be a source oflearning, untapped by top managers in firmswitha national scope. The connectedness of the mapsof the structure of the industry is not related tothe geographic scope of the company (as shownin Table 3). However, the combined measure ofcomplexity of the CEOs' cognitive maps of thestructure (comprehensiveness x connectedness)is weakly related to the geographic scope of thefirm (under the 0.10 level), thus supportingHypothesis 2-1. In firms with an internationalgeographic scope, the CEOs' cognitive maps ofthe structure of the industry are more complex


16/22

CEOs' Cognitive Maps 451

Table3.

Geographicscopeofthefirmand

complexityofthemapsofthe

structureofthe

industry

Dependent

variable

Standardized

comprehensiveness/

Standardized

connectedness/

Standardized

complexity/structure

structure

structure

(i.e.nboflinks)

Factor

Geographicscope

Geographicscope

Geographicscope

Covariate

Lengthofthe

interview

Lengthofthe

interview

Lengthofthe

interview

Maineffect

F=5.55*

sig:0.01F=0.92

sig:0.41F=2.79t

sig:0.08

Covariate

F=1.44

sig:0.24F=1.10

sig:0.30F=0.17

sig:0.69

MultipleR2

R2=0.36

R2=

0.12

2=

0.21

nbofcases

Mean

S.D.nbofcases

Mean

S.D.nbofcases

Mean

S.D.

G1:

National

(scopeofthe

8

0.80

0.17

8

1.07

0.51

8

0.82

0.23

firm)2:

Nationalandafew

9

1.06

0.23

9

1.09

0.26

9

1.17

0.38

foreign

countries

G3:

International

(several

9

1.12

0.22

9

0.86

0.22

9

1

0.22

major

foreign

countries)

Group3>

Group1

Group2>

Group1

(underthe0.05level)

Duncan's

multiplerangetest

Group2>

Group1

N.S.

Group3>

Group1

(underthe0.05level)

(underthe0.10level)

*Significantunderthe0.05level.

'Significantunderthe0.10level.


17/22


18/22

CEOs' Cognitive Maps 453

Table4.

Linkswith

(foreign)

'parent'

companiesand

complexityofthemapsofthe

dynamicsofthe

industry

Dependent

variable

Standardized

comprehensiveness!

Standardized

connectedness/

Standardized

complexity/dynamics

dynamics

dynamics

(i.e.

largest

submap)

Factor

Linkswith

'parent'

companies

Linkswith

'parent'

companies

Linkswith

'parent'

companies

Covariate

Lengthofthe

interview

Lengthofthe

interview

Lengthofthe

interview

Maineffect

F=1.14

sig:0.34F=2.58t

sig:0.09F=3.16t

sig

0.06

Covariate

F=1.79

sig:0.19F=0.00

sig:0.96F=0.92

sig

0.35

MultipleR2

R2=0.16

R2=

0.19

R2=0.25

nbofcases

Mean

S.D.

nbofcases

Mean

S.D.

nbofcases

Mean

S.D.

Gl

Independent

11

0.91

0.21

11

0.91

0.26

11

0.81

0.29

G2:

Belongstoagroup

8

1.06

0.19

8

0.98

0.16

8

1.05

0.31

fromsame

country

G3:

Belongstoa

7

1.07

0.31

7

1.15

0.20

7

1.24

0.41

multinational

(foreign

country)Duncan's

multiplerange

N.S.

Group3>

Group1

Group3>

Group1

test

(underthe0.05level)

(underthe0.05level)

'Significantunderthe0.10level.


19/22

454 R. Calori, G. Johnson and P. Sarnincomplexity of the understanding of the dynamicsis not affected. In Piaget's terms (Piaget, 1937)there is 'assimilation' but not necessarily 'accom-modation' of the mental schemes; there is anincrease in the quantity of knowledge. Theinteractions with other managers from foreignparent companies are sources of real confron-tations of schemas of the dynamics of theenvironment. The quantity of knowledge maynot increase much but the reasoning is enrichedwith new concepts and links explaining thedynamics of the business. In Piaget's terms thereare 'assimilation' and 'accommodation.'The practical implications suggested by thestudy underscore the importance of managementdevelopment in accelerating the learning processwhen executives make major career changes:from a focused firm to a diversified firm, froma business unit with a national scope to one withan international scope, or from an independentcompany to a multinational corporation. Inno-vative training methods may be helpful; forinstance, cognitive mapping can be used to assessindividual learning over time and individualcognitive maps can be enriched by comparisonswith others, thus stimulating reciprocal learning.Basically, in multinational corporations, careerdevelopment should provide a variety of experi-ence in order to develop the cognitive complexityof chief executives who are to play an integrativerole within the top management team.

The results and interpretations should be takenas indicative, bearing in mind the limitations ofthe study already discussed; and in particularour inability to control for individual factorswhich may affect managers' cognition. Weacknowledge that, regardless of the scope of thecompany, a CEO who has held positions inseveral countries in his or her career may havea broader view of the market, and consider morecountries; or a manager with experience ofmultiple functions may have a more comprehen-sive view of the dynamics of the business interms of markets, technologies, human andfinancial resources. The cognitive style of theindividual may also be related to cognitivecomplexity: leaders ('intuitives' and 'feelers' inthe terms used by Hurst et al., 1989) may haveless concepts in their maps than managers('thinkers'), and may be more exclusively focusedon their vision (Noel, 1989). A study based ona larger sample of CEOs would be needed in

order to control for individual factors. Such astudy could also assess if some of the statisticaltests which failed here gain significance in alarger sample, when controlling for individualfactors.The dilemma between technical reliability ofinterview procedures and acceptability by apredetermined sample of chief executives stillhas to be resolved. One way to resolve it wouldbe to focus further studies on one dimension ofthe environment-perceptions of the structureor perceptions of the dynamics-, and to usesimplified forms of existing techniques-forinstance full context repertory grid techniques(Fransella and Bannister, 1977) could be appliedin categorizing competitors and market segments

in the environment.We have not argued that more complexcognitive maps will lead CEOs to superiorperformance. A top manager may utilize a fewconcepts, but they may be the most relevantones. Moreover, managerial work is not onlythinking, it also has to do with feelings andaction. We have however argued that CEOs'cognitive complexity should match the level ofcomplexity of the environment that they areconfronted with, in particular because they haveto play an integrative role within the topmanagement team. The present research hasonly shown relationships between the complexityof the environment mediated by the scope of thefirm and the cognitive complexity of the CEO.As a first step in studying the relationshipsbetween cognitive complexity and performance,it suggests that the scope of the organizationmay have a moderating effect on the relationship.

ACKNOWLEDGEMENTSWe wish to thank Ari Ginsberg and the twoanonymous referees from SMJ for their helpfulcomments on the earlier drafts of this paper.

REFERENCESAnderson, J. (1983). The Architecture of Cognition.Harvard University Press, Cambridge MA.Argyris, C. and D. Schon (1978). OrganizationalLearning: A Theoryof Action Perspective. Addison-Wesley, Reading, MA.


20/22

CEOs' Cognitive Maps 455Ashby, W. R. (1956). Introduction to Cybernetics.Chapman and Hall, London.Ashby, W. R. (1958). 'General systems theory as anew discipline', General Systems Yearbook, Vol.III. pp.1-17.Baddeley, A. (1983). Your Memory: A User's Guide.

Penguin, Middlesex.Bartlett, C. A. and S. Ghoshal (1989). ManagingAcross Borders: The TransnationalSolution. Hutch-inson Business Books, London.Bartunek, J. M., J. R. Gordon and R. P. Weathersby(1983). 'Developing "complicated" understandingin administrators', Academy of Management Review,8(2), pp. 273-284.Beer, S. (1966). Decision and Control. John Wiley,New York.Bennis, W. and B. Nanus (1985). Leaders: TheStrategies for Taking Charge. Harper and Row,New York.Bougon, M. (1983). 'Uncovering cognitive maps: The"Self Q" technique". In G. Morgan (ed.), BeyondMethod: A Study of Organizational Research Strat-egies. Sage, New York, pp. 173-188.Bougon, M. (1992). 'Congregate cognitive maps': Aunified dynamic theory of organization and strat-egy', Journal of Management Studies, 29(3),pp. 369-389.Bower, J. L. (1972). Managing the Resource AllocationProcess. Irwin, Homewood IL.Brown, S. M. (1992). 'Cognitive mapping and repertorygrids for qualitative survey research: Some compara-tive observations', Journal of Management Studies,29(3), pp. 287-307.Chandler, A. D. (1962). Strategyand Structure. M.I.T.Press, Cambridge MA.Cossette, P. and M. Audet (1992). 'Mapping ofan idiosyncratic schema', Journal of ManagementStudies, 29(3), pp. 325-347.Cray, D., G. Mallory, R. Butler, D. Hickson and D.Wilson (1991). 'Explaining decision processes',Journal of Management Studies, 28(3), pp. 227-251.Crossan, M. M. (1991). 'Organization learning: Socio-cognitive model of strategic management', unpub-lished doctoral dissertation, University of WesternOntario.Daft, R. L. and K. E. Weick (1984). 'Towards a

model of organizations as interpretation systems',Academy of Management Review, 9, pp. 284-295.Dearborn, D. C. and H. A. Simon (1958). 'Selectiveperception: A note on the departmental identifi-cations of executives', Sociometry, 21, pp. 140-144.Dill, W. R. (1962). 'The impact of environment onorganizational development'. In S. Mailicks, andE. Van Ness, (eds.), Concepts and Issues inAdministrative Behavior. Prentice-Hall, EnglewoodCliffs, NJ, pp. 94-109.Eden, C. and F. Ackermann (1992). 'The analysis ofcause maps', Journal of Management Studies, 29(3),pp. 309-324.Eiseman, J. (1978). 'Reconciling incompatible po-sitions', Journal of Applied Behavioral Science, 14,pp. 133-150.Fiske, S. T. and S. E. Taylor (1984). Social Cognition.Addison Wesley, Reading MA.

Ford, J. D. and D. A. Baucus (1987). 'Organizationaladaptation to performance downturns: An interpre-tation-based perspective', Academy of ManagementReview, 12(2), pp. 366-380.Fransella, F. and D. Bannister (1977). A Manual forRepertory Grid Technique. Academic Press, NewYork.Ginsberg, A. (1989). 'Construing the business port-folio: A cognitive model of diversification', Journalof Management Studies, 26, pp. 417-438.Ginsberg, A. (1990). 'Connecting diversification toperformance: A sociocognitive approach', Academyof Management Review, 15, pp. 514-535.Greiner, L. E. (1986). 'Top management politics andorganizational change'. In S. Srivastva et al. (eds.),Executive Power. Jossey Bass, San Francisco, CA,pp. 155-177.Grinyer, P. H. and J. C. Spender (1979). 'Recipes,crises and adaptation in mature business', Inter-national Studies of Management and Organization,IX, pp. 113-123.Hickner, Y. E. R. (1991). 'Integrated complexityand profitability', Working Paper, Case WesternReserve University, Cleveland, OH. Presented atthe Academy of Management Conference, Miami,FL.Haley, U. C. V. and S. A. Stumpf (1989). 'Cognitivetrails in strategic decision-making: Linking theoriesof personalities and cognitions', Journal of Manage-ment Studies, 26(5), pp. 477-497.Hambrick, D. C. and P. A. Mason (1984). 'Upperechelons: The organization as a reflection of itstop managers', Academy of Management Review,9, pp. 195-206.Hedberg, B. (1981). 'How organizations learn andunlearn'. In P. Nystrom and W. Starbuck (eds.),Handbook of Organizational Design, Vol. 1. OxfordUniversity Press, Oxford, pp. 3-27.Hedberg, B. and S. Jonsson (Summer 1977). 'Strategyformulation as a discontinuous process', Inter-national Studies of Management and Organization,New York, pp. 88-109.Hedberg, B. and S. Jonsson (1978). 'Designing semi-confusing information systems for organizations inchanging environments', Accounting, Organizationand Society, 3, pp. 47-64.

Hinkle, D. (1965). 'The change of personal constructsfrom the viewpoint of a theory of constructimplications', unpublished Ph.D. thesis, Ohio StateUniversity.Hitt, M. A. and B. B. Tyler (1991). 'Strategic decisionmodels: Integrating different perspectives', StrategicManagement Journal, 12(5), pp 327-352.Hogarth, R. M. (1980). Judgment and Choice: ThePsychology of Decision. Wiley, Chichester.Holsti, 0. R. (1969). Content Analysis for the SocialSciences and Humanities. Addison Wesley, ReadingMA.Huff, A. S. (1982). 'Industry influences on strategyreformulation', Strategy ManagementJournal, 3(2),pp. 119-131.Huff, A. S. (ed.) (1990). Mapping Strategic Thought.John Wiley, Chichester.Huff, A. S. and K. E. Fletcher (1990). 'Conclusion:


21/22

456 R. Calori, G. Johnson and P. SarninKey mapping decisions'. In A. S.Huff (ed.),Mapping Strategic Thought. John Wiley, Chichester,pp. 403-412.Hurst, D. K., J. C. Rush and R. E. White (1989). 'Topmanagement teams and organizational renewal',Strategic Management Journal, Summer SpecialIssue, 10, pp. 87-105.Johnson, G. (1988). 'Re-thinking incrementalism',Strategic Management Journal, 9 (1), pp. 75-91.Kelly, G. A. (1955). The Psychology of PersonalConstructs (Vols. 1 and 2). Norton, New York.Kiesler, S. and L. Sproull (1982). 'Managerial responseto changing environments: Perspectives on problemsensing from social cognition', Administrative Sci-ence Quarterly, 27, pp. 548-570.Langfield-Smith, K. and G. Lewis (1989). 'Mappingcognitive structures: A pilot study to develop aresearch method', Working Paper no. 14, TheUniversity of Melbourne, Graduate School ofManagement, Melbourne.Learned, E.P., C. R. Christensen and K. R. Andrews(1961). Problems and General Management-Busi-ness Policy. Irwin, Homewood IL.Lenz, R. T. and J. L. Engledow (1986). 'Environmentalanalysis: The applicability of current theory',Strategic Management Journal, 7(4), pp. 329-346.Lyles, M. and I. I. Mitroff (1981). 'Organizationproblem formulation: An 1980 empirical study',Administrative Science Quarterly, 25, pp. 102-119.Lyles, M. A. and C. R. Schwenk (1992). 'Topmanagement strategy and organizational knowledgestructures', Journal of Management Studies, 29(2),pp. 155-174.

March, J. and H. Simon (1958). Organizations. Wiley,New York.Martin, J. (1982). 'Stories and scripts in organizationalsettings.' In A. H. Hastorf and A. Isen (eds.),Cognitive Social Psychology. Elsevier North-Hol-land, New York, pp. 255-305.Mason, R. 0. and I.I. Mitroff (1981). ChallengingStrategic Planning Assumptions. Wiley, New York.Melin, L. (1992). 'Internationalization as a strategyprocess', Strategic Management Journal, WinterSpecial Issue, 13, pp. 99-118.Meyer, A. (1982). 'Adapting to environmental jolts',Administrative Science Quarterly, 27, pp. 515-537.Mintzberg, H., 0. Raisinghani and A. Theoret (1976).'The structure of unstructured decision processes',Administrative Science Quarterly, 21, pp. 246-275.Mintzberg, H. (1987). 'The strategy concept I: FiveP's for strategy', California Management Review,30(1), pp. 11-24.Noel, A. (1989). 'Strategic cores and magnificentobsessions: Discovering strategy formation throughdaily activities of CEOs', Strategic ManagementJournal, Summer Special Issue, 10, pp. 33-49.Piaget, J. (1937). La Construction du Reel chez1'Enfant. Delachaux et Niestle, Neuchatel et Paris.Piaget, J., P. Mounoud and J. P. Bronckart (1987).

Psychologie. Gallimard, Paris.Porac, J. F., H. Thomas and B. Emme (1987).'Knowing the competition: The mental models of

retailing strategists'. In G. Johnson (ed.), BusinessStrategy and Retailing. Wiley, New York, pp. 59-79.Porter, M. E. (1980). Competitive Strategy, Techniquesfor Analyzing Industries and Competitors'. FreePress, New York.Porter, M. E. (1985). Competitive Advantage. FreePress, New York.Porter, M. E. (ed.) (1986). Competition in GlobalIndustries. Harvard Business School Press, Boston,MA.Prahalad, C. K. and R. A. Bettis (1986). 'Thedominant logic: A new linkage between diversityand performance', Strategic Management Journal,7(6), pp. 485-501.Prahalad, C. K. and Y. Doz (1987). The MultinationalMission, Balancing Local Demands and GlobalVision. Free Press, New York.Reger, R. K. (1990). 'Managerial thought structuresand competitive positioning'. In A. S. Huff, (ed.),Mapping Strategic Thought. John Wiley, Chichester,pp. 71-88.Reger, R. K. and A. S. Huff (1993). 'Strategic groups:A cognitive perspective', Strategic ManagementJournal, 14(2), pp. 103-124.Schank, R. C. and R. P. Abelson (1977). Scripts,Plans, Goals and Understanding: An Inquiry intoHuman Knowledge Systems. Erlbaum, Hillsdale,NJ.Schrank, R. (1978). Ten Thousand Working Days.M.I.T. Press, Cambridge, MA.Schroder, H., M. Driver and S. Streufert (1967).Human Information Processing. Free Press, NewYork.Schwenk, C. R. (1984). 'Cognitive simplificationprocesses in strategic decision-making', StrategicManagement Journal, 5(2), pp. 111-128.Schwenk, C. R. (1988). 'The cognitive perspective onstrategic decision making', Journal of ManagementStudies, 25(1), pp. 41-55.Shrivastava, P. and G. Lim (1984). 'Alternativeapproaches to strategic analysis of environments',working paper, New York University.Simon, H. A. (1957). Models of Man. Wiley, NewYork.Simon, H. A. (1976). Administrative Behavior, (4thed.). Free Press, New York.Smircich, L. and C. I. Stubbart (1985). 'Strategicmanagement in an enacted world', Academy ofManagement Review, 10, pp. 724-736.Spender, J. C. (1989). Industry Recipes: TheNature andSources of Managerial Judgement. Basil Blackwell,Oxford.Stabell, C B.(1978). 'Integrative complexity of infor-mation environment perception and informationuse: An empirical investigation', OrganizationalBehavior and Human Performance, 22(1),pp. 116-142.Streufert, S., S. C. Streufert and C. H. Castore (1968).'Leadership in negotiations and the complexity of

conceptual structure', Journal of Applied Psy-chology, 52, pp. 218-223.Streufert, S. (1972). 'Success and response rate in


22/22

Documents

1ceo Cognitive Map