Chief executive scanning emphases, environmental dynamism, and manufacturing firm performance

Strategic Management JournalStrat. Mgmt. J., 24: 725–744 (2003)

Published online 23 May 2003 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/smj.335

CHIEF EXECUTIVE SCANNING EMPHASES,ENVIRONMENTAL DYNAMISM, ANDMANUFACTURING FIRM PERFORMANCE

VINAY K. GARG,1 BRUCE A. WALTERS2* and RICHARD L. PRIEM3

1 College of Business Administration, Southwest Missouri State University, Spring-field, Missouri, U.S.A.2 College of Administration and Business, Louisiana Tech University, Ruston,Louisiana, U.S.A.3 School of Business Administration, University of Wisconsin-Milwaukee, Milwaukee,Wisconsin, U.S.A.

Chief executives must allocate their scarce time for scanning efforts among relevant domains oftheir firms’ external environment and their firms’ internal circumstances. We argue that high-performing CEOs vary their relative scanning emphases on different domains according to thelevel of dynamism they perceive in their external environments. The concepts of dominant logicand sector importance were used to develop predictions about which external domains andwhich internal domains should receive relatively more or less scanning emphasis in externalenvironments that, overall, are more dynamic or more stable. A field survey of 105 single-business manufacturing firms evaluated CEOs’ scanning emphases and firm performance. Resultsindicated that, for dynamic external environments, relatively more CEO attention to the tasksectors of the external environment and to innovation-related internal functions was associatedwith high performance. In stable external environments, however, simultaneously increasedscanning of the general sectors in the external environment and efficiency-related internalfunctions produced higher performance. These relationships were strongest between relativescanning emphases among domains and sales growth. We discuss the implications of these resultsfor researchers and practitioners. Copyright 2003 John Wiley & Sons, Ltd.

Chief executives must selectively allocate theirattention: their time is scarce; myriad external andinternal data are available to them; and these datatypically represent complex, hard-to-interpret phe-nomena. CEOs must therefore focus on what theyperceive to be key subsets of the available data,and by necessity exclude some other, potentiallyimportant data sources (Hambrick, 1981; Prahalad

Key words: CEO; scanning; dynamism; manufacturing;performance*Correspondence to: Bruce A. Walters, College of BusinessAdministration, Louisiana Tech University, PO Box 10318, Rus-ton, LA 71272, U.S.A.

and Bettis, 1986). This ‘scanning selection’ task ismade even more difficult because some inconse-quential data sources may be readily available oreven stridently seek attention, while other, morecritical sources may be elusive, subtle, or hiddenfrom view. In spite of these challenges, effectivescanning emphasis remains a prerequisite to suc-cessful organizational adaptation. As Hambrick hasnoted, an ‘organization’s executives can only acton those phenomena to which their attention isdrawn’ (Hambrick, 1981: 299).

Effective scanning underlies the sound executivejudgment and choices that are essential for strategic

Copyright 2003 John Wiley & Sons, Ltd. Received 19 March 1998Final revision received 3 March 2003

726 V. K. Garg, B. A. Walters and R. L. Priem

success (Child, 1972, 1997; Priem and Cycyota,2001). Thus, scanning may represent a ‘dynamiccapability’ for the firm (Eisenhardt and Mar-tin, 2000). Moreover, executives are likely tohave greater discretion over actions like scan-ning, which involve their own, individual activi-ties, than they have over organization-wide activ-ities like strategy implementation. Therefore, theyare able to effect change in their scanning pat-terns when they believe it is warranted, as mightbe the case when they believe their environmenthas changed. Finally, executives must consciouslyselect and accurately scan the appropriate envi-ronmental sectors before they can hope to initiateeffective actions to align their organization withthe demands of those sectors. This choice, of whereto allocate limited time and attention, is the topicof our study.

Surprisingly, executive scanning emphasis hasreceived less empirical attention from researchersthan have some other areas—such as planning,analysis, strategy or decision making—for whichscanning is an influential antecedent. Relatively lit-tle is known about whether a firm’s context mayinfluence the particular sectors of the external envi-ronment to which an executive’s attention may bedrawn, or about how scanning emphasis on par-ticular sectors may be associated with firm perfor-mance. An additional factor that limits our knowl-edge of executive scanning has been the nearlyexclusive emphasis of research to date on the exter-nal environment (e.g., Aguilar, 1967; Daft, Sor-munen, and Parks, 1988; El Sawy, 1985). A firm’sinternal circumstances also produce important andchangeable data that compete for executive timeand attention, yet this aspect of executive scanninghas been relatively ignored (Bluedorn et al., 1994;Eisenhardt, 1989). The role of firms’ internal capa-bilities in building sustainable advantage, givenmore balanced treatment in early theory (e.g.,Andrews, 1980; Ansoff, 1965), has recently beenreemphasized through the ‘resource-based view’of strategy (Barney, 1991). If a firm’s capabili-ties are indeed important to performance, executivescanning must produce astute comprehension ofchanges in both the external environment and theinternal circumstances of a firm before effectiveadaptation can occur, and both must be consideredin scanning research.

Given that: (1) CEOs must make trade-offsregarding which data to attend; (2) executive scan-ning is an influential antecedent to subsequent

action; and (3) effective scanning involves atten-tion to aspects of both the firm’s external environ-ment and its internal circumstances; it follows thatexecutive scanning emphasis—that is, the relativescanning effort given to various external and inter-nal domains—may warrant more attention in itsown right as a contributor to successful firm adap-tation. Our research follows this line of thinking byasking: Which sectors of the external environment,and which aspects of the internal circumstances,should receive relatively more or less scanningemphasis by CEOs of manufacturing firms facingdifferent competitive environments?

LITERATURE REVIEW ANDHYPOTHESES

Selective attention

Executives are ‘sophisticated information seek-ers’ (Boyd and Fulk, 1996: 2) who neverthelessare constrained by bounded rationality (Cyert andMarch, 1963). Because the magnitude of the scan-ning task is beyond any individual’s informationprocessing capability, top executives must set pri-orities in order to cope. Even with this selectiv-ity, however, efforts to monitor solely the externalenvironment can account for as much as one quar-ter of top executives’ time (Hambrick, 1981).

Thus, CEOs’ cognitive limits (Simon, 1957)force them to pay selective attention to key criteriain making strategic decisions. March and Simon(1958) further proposed that variables which arelargely within the control of problem-solving indi-viduals or organizational units are considered first.If a satisfactory program of activity is not discov-ered based on these variables, then attention will beshifted to variables not under direct control. More-over, Thomas, Clark, and Gioia (1993) found thathigher levels of information use during scanninglead to a heightened sense of controllability andpositive gain regarding these areas, and Sutcliffe(1994) argued that more intense and frequent scan-ning enhances the recognition of environmentalchanges. Taken together, these findings imply that:(1) CEOs’ cognitive limits mandate that they beselective in the scanning task; (2) higher informa-tion use increases the sense of controllability andaccuracy regarding the environment; and (3) thefocus of information use should be directed pri-marily to areas deemed important given the com-petitive environment in which the firm finds itself.

Copyright 2003 John Wiley & Sons, Ltd. Strat. Mgmt. J., 24: 725–744 (2003)

Chief Executive Scanning Emphases 727

Scanning the external environment

Empirical results have supported the idea thatselectivity in scanning contributes to firm perfor-mance (e.g., Boyd and Fulk, 1996; Daft et al.,1988). CEOs must invest their ‘scarce execu-tive time’ (Boyd and Fulk, 1996) in effectivescanning activities that will ‘pay off’ (Fredrick-son and Mitchell, 1984). Further, sectors of theexternal environment differ in importance anduncertainty (Daft et al., 1988), and firms can ben-efit from emphasis on the ‘right’ sectors. Anunderlying theoretical explanation has yet to beadvanced, however, specifying the combinations ofsectors that should be emphasized in differing con-texts. Questions remain concerning which sectorsshould be given most attention, when, and why.Concerning the external environment, Daft et al.(1988) noted that executives could overcome timeand information-processing capacity limitations byfocusing their scanning on specific sectors ratherthan scanning broadly. Following from the work ofBourgeois (1980) and Dill (1958), they theorizedthat the task environment (customer, competitor,and technological sectors) changes more rapidlyand is therefore perceived as more important thanthe general environment (social, economic, andregulatory sectors).

Scanning the internal environment

The early normative literature on strategic planningplaced as much emphasis on the internal circum-stances of the firm as it did on the external envi-ronment (Learned et al., 1965). ‘The central tenetin strategic management is that a match betweenenvironmental conditions and organizational capa-bilities and resources is critical to performance,and that a strategist’s job is to find or createthis match’ (Bourgeois, 1985: 548). Andrews sim-ilarly recognized that both opportunity and com-petence must be accurately identified for strate-gic success. In noting the challenges of internalscanning, he argued that: ‘The insight required toidentify the essential strength justifying new ven-tures does not come naturally. Its cultivation canprobably be helped by recognition of the needfor analysis’ (Andrews, 1980: 67). More recently,Goodstein, Pfeiffer, and Nolan (1991) have alsoadvocated monitoring the organization’s internalcircumstances as part of the strategic planning pro-cess.

The ‘resource-based view’ has reemphasizedthe importance of analyzing a firm’s internalstrengths and weaknesses during the strategy-making process (Barney, 1991; Wernerfelt, 1984).Work focusing on the less tangible, knowledge-based resources has suggested that the capabili-ties generated by such resources: are subtle andhard to understand; involve elusive talents thatare difficult to link to results; can change quiterapidly; and are important to competitive advan-tage (e.g., Eisenhardt and Martin, 2000; Lippmanand Rumelt, 1982; Miller and Shamsie, 1996).These factors imply, first, that an acute compre-hension of firm capabilities is likely necessaryfor successful capabilities–environment matching,and second, that due to internal change executivesmust repeatedly scan internally to maintain theirunderstanding of the capabilities of their firms forcontinued, effective adaptation. We therefore arguethat scanning the internal domains is necessary,in conjunction with scanning the external environ-ment, for effective organizational adaptation. Anadaptive strategic fit requires relating the firm’sstrengths and weaknesses to specific opportunitiesand threats embedded in the external environment;thus, simultaneous scanning of the firm’s externalenvironment and internal circumstances is neces-sary. Most scanning research to date, however, hasexamined only the external scanning aspects ofthis theoretical equation, resulting in underspeci-fied empirical models.

Environmental dynamism

Dynamism describes the rate and the unpredictabil-ity of change in a firm’s external environment(Dess and Beard, 1984), and is particularly impor-tant because of its influence on relationships bet-ween a variety of firm-level constructs and firmperformance. Examples include: organizationalstructure (e.g., Burns and Stalker, 1961), business-level strategy (e.g., Miller, 1988), and strategy-making processes (see Rajagopalan, Rasheed, andDatta, 1993, for a review, and Priem, Rasheed, andKotulic, 1995, for an example). Milliken (1987)emphasized the need for expending greater scan-ning time and effort, overall, under dynamic envi-ronments. Daft et al. found that CEOs of high-performing firms ‘tailor scanning more closely toperceived uncertainty. Scanning efforts are focusedmore directly on areas of strategic uncertainty and



information need’ (Daft et al., 1988: 134). More-over, considerable empirical work has found posi-tive relationships between the simple level of scan-ning efforts and performance among firms oper-ating in variable environments (e.g., Daft et al.,1988; Glick, Miller, and Huber, 1993; Priem et al.,1995).

Our study addressed the relative scanning em-phasis given to different sectors within the firm’sinternal capabilities and external environment,given a particular level of scanning effort, com-pared across firms. That is, does the relativeemphasis among sectors affect performance evenafter considering the effects of the overall levelof time and effort expended? In the next section,the concepts of dominant logic and sector impor-tance are used to develop predictions about whichinternal sectors and which external sectors shouldreceive relatively more or less CEO scanningemphasis in external environments that, overall,are more stable or more dynamic.

The scanning–environment match

The scarcity of executive time indicates that CEOswill focus their scanning efforts on a limited num-ber of sectors. Given this propensity to focus scan-ning effort, however, the question remains: Whatdetermines the scanning emphases that should bepursued by CEOs? Research has shown environ-mental dynamism to be related to many impor-tant strategy variables. We argue in the follow-ing paragraphs: (1) that more dynamic environ-ments require more innovation; (2) that innovationis associated with particular aspects of the exter-nal environment and particular internal functionsof the firm; and (3) that CEOs whose dominantlogic emphasizes innovation will find innovation-associated external aspects and internal functionsparticularly important to scan. We also make simi-larly structured arguments for stable external envi-ronments and efficiency. In each part, we tierelative scanning emphasis on particular externalaspects and particular internal functions to per-formance for firms facing more dynamic or morestable external environments. The concept of dom-inant logic is crucial to our arguments in the nextsections. ‘Dominant logic’ among a firm’s topmanagers is defined as ‘the way in which man-agers conceptualize the business and make criticalresource allocation decisions’ (Prahalad and Bettis,1986: 490). We now compare effective scanning

emphases for two different dominant logics: onestressing dynamism and innovation, and anotherstressing stability and efficiency.

The dominant logic: dynamism and innovation

Strategy research suggests that firms facing dy-namic environments must innovate to succeed.Miles and Snow’s (1978) case studies, for exam-ple, show that ‘prospector’ firms prosper by con-fronting uncertainty through product innovation.Firms that do not innovate in dynamic environ-ments fall behind the frequent changes in productsand practices common in such environments, andare likely to lose sales (Duncan, 1972; Miller,1988). Miles and Snow’s (1978) ‘reactors,’ forexample, attempt to reduce uncertainty by ignor-ing their environments and neglecting innovation.These types of firms often perform poorly enoughto face the danger of perishing, as one might expectfrom Bourgeois’ argument that ‘firms should onlyreduce uncertainty under stable environmental con-ditions and that uncertainty may be functionalin volatile environments’ (Bourgeois, 1985: 570,emphasis in original). Thus, firms that confrontuncertainty where it exists, via innovation, typ-ically outperform those that ignore its presence.Effective innovation, however, requires scanningfocus on appropriate external and internal domains.

Innovation is typically associated with the tasksectors of the external environment, since thesesectors offer more immediate opportunities forgoal attainment and are likely to change morerapidly than are general sectors (Bourgeois, 1980;Daft et al., 1988; Dill, 1958). Product innovationfrequently: is driven by changes in customer tastes;provokes reactions from competitors; and requiresnew technology for implementation (Hofer andSchendel, 1978; Miles and Snow, 1978; Millerand Friesen, 1984; Zaltman, Duncan, and Holbeck,1973). Innovation helps firms to reap pioneer-ing advantages in environments characterized byintense competition, exponential advances in tech-nology, and frequently shifting customer require-ments (Kessler and Chakrabarti, 1996).

The need for particular scanning emphases byCEOs pursuing innovation arises also for inter-nal scanning, because just as external informationis abundant, so is internal information. Further,just as scanning emphasis on particular sectors ofthe external environment is necessary for firms



facing dynamic environments, so may be scan-ning particular domains within the organization.When developing and maintaining organizationalcompetencies in key technologies through inno-vation, for example, the market research, prod-uct R&D, and basic engineering internal functionsmay be seen as particularly important by CEOs.Eisenhardt’s (1989) inductive study provides ananecdotal example of internal domains that drawCEOs’ attention in dynamic environments: the firstemployee hired by the CEO of a firm facing a high-velocity environment was a manager whose jobwas to track new product development projects.

These arguments suggest that the dominant logic(Prahalad and Bettis, 1986) of effective CEOswhose firms are facing dynamic external envi-ronments will likely revolve around the externalenvironment sectors and the internal capabilitiesmost closely associated with successful innovation.Managers of successful firms in dynamic environ-ments are thus likely to conceptualize their busi-ness as involving customer-, technology-, and/orcompetitor-driven innovation, and to considerinternal capabilities in market research, productR&D, and basic engineering as critical to success.

Innovation-centric dominant logic will likelymake task sectors of the external environmentand innovation-related internal capabilities moresalient to CEOs of firms facing dynamic environ-ments. Ratneshwar et al. (1997) found that highsalience is related to selective attention, with thosebenefits that are most salient receiving greatestattention. One might similarly expect that thosesectors of the external and internal environmentsthat are most closely associated with the dominantlogic in a firm will have highest salience for theCEO and thus receive more attention than wouldless salient sectors.

In sum, these arguments suggest that, for dy-namic environments requiring innovation, effectiveCEOs must draw frequently from their knowl-edge of the external task environment and inter-nal innovation functions. Scanning emphasis onexternal task sectors is advantageous in dynamicenvironments because: (a) scanning the compet-itive environment helps executives keep trackof, and react to or preempt, competitors’ ini-tiatives; (b) scanning the technological environ-ment sensitizes them to available solutions thataddress internal weaknesses or external opportuni-ties; and (c) scanning the customer sector helps inearly identification of changes in customer tastes.

In dynamic environments, Eisenhardt’s medium-grained research suggests that executives: makeuse of real-time information, ‘especially on afirm’s competitive environment and operations’(Eisenhardt, 1989: 549); consider multiple alterna-tives simultaneously; and integrate decisions withone another and with tactical plans as meansto arrive at comprehensive and speedy decisionsthat lead to higher performance. Only simul-taneous scanning of external task and internalinnovation domains provides the ‘raw material’(i.e., the knowledge necessary for matching keyenvironmental conditions with the right organi-zational capabilities) from which executives indynamic environments can make sound decisions.Knowledge of the appropriate external environ-ment domains without corresponding knowledgeof the appropriate internal capabilities, or viceversa, will not allow effective organizational adap-tation. Thus:

Hypothesis 1:As perceived environmental dy-namism increases, and as scanning emphasis onthe internal functions dealing with innovationincreases, an increased CEO scanning emphasison the external task environment is associatedwith higher firm performance.

The dominant logic: stability and efficiency

Although crucial in dynamic environments, strongefforts toward product innovation may be super-fluous and inefficient in more stable environments(Duncan, 1972; Miller, 1988). This is because:(a) when competitive pressures are not intense,innovation by a firm may usurp its establishedcompetitive advantages; (b) when technologicalchanges are not rapid, fewer opportunities maybe available to a firm to exploit; and, (c) whencustomer preferences do not change rapidly, prod-uct innovations by a firm may prematurely shortenthe life cycles of its strong products (Kesslerand Chakrabarti, 1996). Instead, spending timeand effort scanning the more remote domainsof the external general environment under stableconditions can open up possibilities for devel-oping long-term, resource-enhancing opportuni-ties, and for identifying major threatening events.Increased scanning emphasis on the general envi-ronment under stable conditions is unlikely todetract from performance, because scanning thetask environment won’t absorb as much executive



time when changes take place slowly. Boyd andFulk (1996) reported anecdotal evidence support-ing these notions. Further, Ansoff’s (1965) ‘aware-ness strategy’ suggested that a firm can gain astrategic advantage during periods of stability bycollecting information on vague, ill-defined (i.e.,more macro) issues.

Similarly, scanning efficiency functions such ascost controls and operating efficiency may takeon greater importance as the external environmentbecomes more stable. Organizations can analyzelong-term, capability-building options for meetingslowly unfolding trends in the general environ-ment. Capital investments that could reduce coststo a particular target level, for example, might opennew avenues for achieving long-lasting compet-itive advantage. The joint exploration of slowlydeveloping opportunities and hard-to-build compe-tencies is possible only in more stable conditions.

These arguments suggest that the dominant logicof firms facing stable external environments willlikely revolve around the general sectors of theexternal environment and efficiency-related inter-nal capabilities. Following the logic outlined forHypothesis 1: (a) the knowledge of stable-en-vironment CEOs is most appropriate when theyview the general sectors of the external envi-ronment and the efficiency-related internal capa-bilities as key to successful adaptation; (b) theimportance of these external and internal domainswill be higher for CEOs facing stable exter-nal environments than for CEOs of firms fac-ing more dynamic environments; (c) CEOs withmore appropriate knowledge will put relativelymore scanning emphasis on the key domains; and(d) simultaneous scanning emphasis on the keyexternal and internal domains is necessary for suc-cessful adaptation. Thus:

Hypothesis 2: As perceived environmental dy-namism decreases, and as scanning emphasison the internal functions dealing with efficiencyincreases, an increased CEO scanning emphasison the external general environment is associ-ated with higher firm performance.

In summary, our discussion of the literature estab-lished that an organization must have the appropri-ate competencies in order to capitalize on externalopportunities. We have further argued that CEOscanning efforts should emphasize the firm’s key

competencies and opportunities, and that the spe-cific domains of scanning attention are contingenton the level of dynamism in the firm’s externalenvironment. In order for the organization to adaptwell when facing a particular level of dynamism,the CEO must match the appropriate scanningemphases in external domains with the appropriateemphases in internal domains. Thus, appropriatescanning of the external environment alone, or ofthe firm’s internal circumstances alone, would notlead to high performance. Finally, ‘emphasis’ doesnot imply that the absolute level of scanning fora particular domain is greater than for another;rather, emphasis requires that a firm pay moreattention to that particular domain, relative to othersectors, when compared to other firms.

These arguments suggest that the ‘scanningselection’ task is a contingent one. A context-specific approach to scanning prioritization maybe necessary for recognition of suitable oppor-tunities to enhance firms’ resources (Pfeffer andSalancik, 1978), build strategic advantage (Duttonand Freedman, 1984), and avoid potential threats(Ansoff, 1975).

METHOD

Sample and procedure

Strong tests of hypotheses involving triple interac-tions, such as ours, are generally very difficult toplan in field studies (McClelland and Judd, 1993).We maximized statistical power through our dataanalysis procedure (discussed later) and throughour sampling plan. We followed Harrigan’s (1983)suggestion that researchers select samples carefullyto maximize the chances of detecting hypothesizedeffects statistically. We established three criteria, apriori, for identifying an appropriate population offirms from among those in a large southwesternmetropolitan area that appeared in the Directory ofManufacturers.

First, to be included in the sample the firm musthave been an independent business rather than asubsidiary, a division of another firm, or a unitof a conglomerate. This criterion ensured that theexecutive’s scanning behavior was not influencedby a parent firm, and therefore controlled for firmautonomy (Child, 1972). Second, the firm musthave been a single-business firm (Rumelt, 1974).This was determined through each firm’s self-classification in the Directory of Manufacturers as



operating in only one 4-digit SIC code, or in two4-digit SICs that could be considered in the sameindustry. This helped to ensure that the responsesof the executives focused only on the environmentof the primary business, rather than on the potentialmultiple environments associated with diversifiedfirms. Finally, firms in the sample must have hadfrom 50 to 99 employees. The minimum of 50employees helped ensure that the business was aconsequential entity. The maximum of 99 employ-ees was established because we anticipated that thescanning emphasis of the CEO would likely havegreater influence on performance for smaller thanfor larger firms (e.g., Miller and Toulouse, 1986).In larger firms the boundaries at which transactionswith the external environment take place are farremoved from the CEO, and the status of the inter-nal environment may be less accurately perceivedby the CEO, because of filtering by multiple lay-ers of managers. Smaller-firm CEOs, on the otherhand, may be more involved in both external trans-actions and internal implementation. Thus, theirscanning would likely have greater influence onfirm outcomes.

The 320 firms that were initially identified asmeeting the three criteria formed the sample pop-ulation. We anticipated a response rate of less than30 percent because of two factors. First, becausethey are located in an area with many universities,some firms may have received numerous previousrequests to participate in research studies. Second,the top executives of smaller firms frequently workat both strategic and operating levels, leaving themwith little time to participate in academic research.Nevertheless, our goal was to have at least 100firms in the final sample. After follow-up mail-ings and phone calls, 116 firms returned usableresponses. Three of these firms employed 500 ormore employees, and therefore were excluded asmuch too large. Thus, our overall response ratewas 35 percent.

The mean size of the firms was 80 employees(S.D. = 37). They manufactured a variety of prod-ucts. The average CEO was 50 years old (S.D. =10.5) and had 21.6 years (S.D. = 12) of industryexperience. Their average tenure with the firm was16 years (S.D. = 11) and in the position of CEOwas 11.6 years (S.D. = 10). The executives hadvaried functional backgrounds, although sales andmanufacturing were somewhat better represented.A third of the CEOs had formal business educa-tion, mostly at the undergraduate level.

Data collection

The CEOs or presidents of the firms identified asmeeting our criteria were sent a letter request-ing participation in the study and a packet con-taining three questionnaires. The letter asked theCEO to complete a questionnaire, and to pass theother questionnaires to others in the firm who wereinvolved in scanning and strategy making. Respon-dents were promised anonymity and a summary ofthe results that would indicate how their firm com-pared to other firms in the sample. The question-naire was pilot tested using three firms that werenot included in the study. This resulted in minorchanges to the questionnaire to increase clarity.

Thirty-four of the responding firms returnedmore than one survey. Follow-up phone calls indi-cated that the additional respondents tended to beprimarily responsible for one major function ratherthan organization-wide concerns. Moreover, overhalf of the CEOs of the single-response firms statedin follow-up calls that they were the only execu-tives at their firm involved in environmental scan-ning and strategic decision making. We thereforeadopted a key informant approach (see Kumar,Stern, and Anderson, 1993; Seidler, 1974), usingdata provided by the CEOs only. We conductedinterrater reliability checks on organizational-levelvariables such as size, environmental dynamism,the overall level of scanning, and performance,however, for those firms where we did have mul-tiple respondents. These reliabilities, ranging from0.69 to 0.88, were satisfactory and validate the useof the key informant approach.

Operationalization of variables

Scanning emphases

We defined a CEO’s scanning emphases as therelative importance the CEO places on informa-tion from differing domains of the external envi-ronment and internal circumstances of the firm.Thus, our first task in operationalizing scanningemphases was to identify relevant domains offirms’ external environment and internal circum-stances. Daft et al. (1988) had identified the ‘task’and ‘general’ environments as relevant domains ofthe external environment. The empirical groupingsthey found for some subsectors, however, were notconsistent with their expectations. Even less empir-ical guidance was available for classifying domainsof the internal environment.



We began with six questionnaire items used tomeasure CEO-perceived importance of the externalenvironment domains, one for each sector iden-tified in Daft et al. (1988) and Aguilar (1967).These included the market environment, tech-nological environment, competitive environment,political/legal environment, economic environ-ment, and sociocultural environment. Likewise, sixitems were used to measure the importance ofthe internal environment domains, one for eachactivity in the value chain believed by Porter(1985) to be critical to competitive advantage.These included market research, product R&D,basic engineering, financial management, cost con-trols, and operational efficiency. Each item askedthe CEO to rate the importance of the sector on a7-point scale ranging from 1 (very important) to 7(not at all important).

For parsimony to reduce the dimensionalityof these CEO data, and because of the relativescarcity of empirical data on CEO-perceived exter-nal and internal domains, we performed explora-tory common factor analyses using oblique rotationon the importance data obtained from the manu-facturing firm CEOs. Oblique rotation was usedbecause there was no theoretical basis for assumingthat the domains would be fully orthogonal. Theexternal environment domains were factor ana-lyzed first, and then the internal domains. We keptthe external and internal domains separated, in sep-arate factor analyses, because the theory of strate-gic adaptation used in the hypothesis developmentsection requires appropriate matching of the exter-nal environment’s characteristics with the internalcapabilities of the firm. This precludes external andinternal areas from falling in the same factor.

Multidimensionality was demonstrated in boththe analyses. Each exhibited convergence on a two-factor solution via complementary methods—suchas different varieties of the eigenvalue criterion,the scree-plot, and interpretability—as advocatedby Kim and Mueller (1978) and Loehlin (1992).In each case, a simpler structure was obtainedfor the two-factor rather than the single-factor orthree-factor solutions. Factor loadings for the two-factor solutions for both the external and internalenvironment domains are shown in Tables 1 and 2,respectively.

For the external environment, the sociocultural,economic, and political/legal sectors loaded on thefirst factor (labeled ‘scanning emphasis on thegeneral environment’). Market, technological, and

Table 1. Factor loadings: external environment items

Item Factor 1 Factor 2

Importance of socioculturalenvironment

0.81 0.11

Importance of economicenvironment

0.68 0.28

Importance of political/legalenvironment

0.54 0.32

Importance of marketenvironment

0.17 0.70

Importance of technologicalenvironment

0.26 0.60

Importance of competitiveenvironment

0.15 0.48

Table 2. Factor loadings: internal environment items

Item Factor 1 Factor 2

Importance of cost controls 0.91 0.15Importance of operational

efficiency0.87 0.16

Importance of product R&D 0.06 0.80Importance of market research 0.30 0.63Importance of financial

management0.51 0.61

Importance of basicengineering

−0.03 0.47

competitive sectors loaded on the second factor(named ‘scanning emphasis on the task environ-ment’). These loadings are consistent with the lit-erature (e.g., Daft et al., 1988) and permit us to testhypotheses involving differential effects of scan-ning emphases on task vs. general domains of theexternal environment.

For the internal environment, cost control andoperating efficiency loaded on the first factor.Product R&D, market research, and basic engi-neering loaded on the second factor. Financialmanagement did not exhibit a simple structure,loading on both factors. Therefore, it was droppedfrom further analyses. We labeled the first factor‘scanning emphasis on efficiency’ and the secondfactor ‘scanning emphasis on innovation.’ Theseloadings are consistent with Porter (1980): firmsthat pay more attention to functions such as costcontrol and operating efficiency typically tend toemphasize efficiency. On the other hand, sensitiv-ity to trends and developments related to basicengineering, technology and markets, via scan-ning of internal functions such as product R&D,



market research, and basic engineering, permitsfirms to pursue innovation. The two internal envi-ronment factors, therefore, allow us to test forthe differential effects of scanning emphasis onefficiency-oriented vs. innovation-oriented capabil-ities of the firm.

Environmental dynamism

We measured environmental dynamism using themulti-item scale of Miller and Droge (1986). ThisLikert-type scale has seen frequent use in thestrategy literature (e.g., Miller, 1988; Priem et al.,1995). Our interest was in the CEOs’ perceptionsof the level of dynamism, because executives acton their perceptions. Further, the study of strat-egy processes such as executive information searchis likely to benefit most from the use of percep-tual measures (Boyd, Dess, and Rasheed, 1993;Boyd and Fulk, 1996). The questions, measuredon a 1–7 scale, are as follows: ‘Our firm mustrarely change its marketing practices to keep upwith the market and competitors’ vs. ‘Our firmmust change its marketing practices extremely fre-quently (e.g., semi-annually)’; ‘The rate at whichproducts/services are getting obsolete in the indus-try is very slow (e.g., basic metal like copper)’vs. ‘The rate of obsolescence is very high as insome fashion goods’; and ‘The production/servicetechnology is not subject to very much change andis well established (e.g., in steel production)’ vs.‘The modes of production/service change often andin a major way (e.g., advanced electronic compo-nents).’

Firm performance

We expected most of the small firms in the sam-ple to be privately held. We therefore believedit unlikely, based on our past experience, thatthe CEOs would be willing to provide detailedaccounting data on firm performance. Therefore,we used subjective, self-reported measures of per-formance. The CEOs were asked to report theirbest subjective estimates of performance comparedto similar firms in their industry on a 5-point scalefor each of the following: after tax return on totalassets, after tax return on total sales, sales growth,and overall performance/success. Measures suchas these have been found to be highly correlatedwith objective measures of firm performance (e.g.,Dess and Robinson, 1984; Robinson and Pearce,

1988; Venkatraman and Ramanujam; 1987). More-over, the literature suggests that subjective mea-sures should be used when interest centers oncapturing the perspective of organizational mem-bers (Duncan, 1972) and when studying manage-rial behavior and decision making (Boyd et al.,1993). Five-year average estimates were elicited.This minimized the influence of short-term per-formance variations, but did in turn require thatthe CEOs had relatively long tenures as CEO.The explicit performance comparisons to similarfirms provided a form of control for differences inperformance due to industry (Dess, Ireland, andHitt, 1990) and strategic group (Hatten, Schen-del, and Cooper, 1978) effects. Lastly, the mul-tiple measures were appropriate considering themultidimensionality of the performance constructs(Cameron, 1978; Chakravarthy, 1986). We found ahigh correlation (0.84) between after tax return ontotal assets and after tax return on total sales and,therefore, combined these two into an ‘accountingreturn’ measure of performance.

Control variables

Consistent with earlier research on smaller firms,we measured firm size using number of employees.The size variable was transformed using the naturallogarithm to obtain a normal distribution, and thetransformed variable was renamed ‘logsize.’ Theanalyses were controlled for logsize because sizeis a frequent correlate of firm performance. In addi-tion, the analyses were controlled for the overalllevel of scanning, so CEOs’ individual predilec-tions toward scanning would not overwhelm theeffects of the relative emphases of their scanning,which is the research interest in this study. Levelof scanning was measured using Miller’s (1987)four-item, 7-point Likert-type scale, which gathersinformation on the extent to which various scan-ning devices are used by the executive’s firm.

Data analysis

Hierarchical regression was used to test the hypoth-eses. To reduce the potential problem of multi-collinearity, we centered variables prior to formingmultiplicative terms (Jaccard, Turrisi, and Wan,1990). Because the interaction terms used to testmoderation frequently are correlated with the termsfrom which they are constructed, ‘it is necessaryto partial the component parts of the product term



from the term itself when evaluating the presenceof a moderated relationship. This is the essence ofthe hierarchical test’ (Jaccard et al., 1990: 24). Ourhypotheses were for ‘double moderation,’ whereinthe levels of two variables together influence thesize of the effect (i.e., the slope) that a third vari-able has on the dependent performance variable.Thus, we tested our hypotheses with triple interac-tion terms, with all main effects and double inter-actions included in the hierarchical process. Onlythe control variables, the scanning domain vari-ables, and environmental dynamism were enteredin the first step of the regression model. Doubleinteractions were added in the second step, andthe appropriate triple interactions were added inthe final step (Bobko, 1995; Darlington, 1990).

Our revised full regression model is:

Y = a1logsize + a2 scanning level + b1X1

+ b2X2 + b3X3 + b4X4 + b5X5 + b6X1X5

+ b7X2X5 + b8X3X5 + b9X4X5 + b10X1X3

+ b11X1X4 + b12X2X3 + b13X2X4 + b14X1X3X5

+ b15X1X4X5 + b16X2X3X5 + b17X2X4X5 (1)

where Y = firm performance, X1 = scanning em-phasis on task environment, X2 = scanning empha-sis on general environment, X3 = scanning empha-sis on innovation, X4 = scanning emphasis on effi-ciency, X5 = perceived environmental dynamism,and a1, a2, b1 to b17 are regression coefficients.

In the above model, only the appropriate tripleinteractions, rather than all possible triple mul-tiplicative terms, have been included. Our over-arching theory argued that higher firm performancewould result when one pre-identified scanning tar-get from the external domain and one pre-identifiedscanning target from the internal domain receivehigher scanning emphasis of CEOs facing eithera more dynamic or a more stable environment.Thus, any multiplicative term that does not includeone target from the external domain, one targetfrom the internal domain, and perceived environ-mental dynamism would be considered inappropri-ate. After eliminating inadmissible combinations,we were left with four possible triple combina-tions. Although our hypotheses include only twoof those four combinations, at this early stage oftheory development and exploratory analysis wedecided to include the remaining two combinationsas implicit controls. By doing so, we are able to

demonstrate the unique effects of our hypothesizedinteractions, over and above any possible effects ofthe triple interactions that are not hypothesized butthat also meet the overarching criteria explainedearlier in this paragraph.

Our multivariate, contingency Hypothesis 1would be supported by a positive, significant tripleinteraction term for the influence of task scanning,innovation scanning, and dynamism together onthe performance measures. Plus, when graphed theinteraction would show that the effect of dynamismon performance is greatest at higher levels of taskscanning and higher levels of innovation scanning.Hypothesis 2 would be supported by a negative,significant triple interaction term for the influ-ence of general scanning, efficiency scanning, anddynamism together on the performance measures.Further, when graphed the interaction would showthat the negative effect of dynamism on perfor-mance is greatest at higher levels of general scan-ning and higher levels of efficiency scanning.

The power to detect interactions is generallylow in field studies, due to factors including morenoise, less control over measurement errors, thenature of interactions (i.e., often ordinal), andthe potential nonlinearity of bivariate relation-ships involving the dependent variable (McClel-land and Judd, 1993). Given that: (1) our hypothe-ses involve triple interactions; (2) the theory lead-ing up to our hypotheses is well entrenched in thestrategy literature; (3) there is no prior evidencethat the effect sizes will be large, because sim-ilar hypotheses have not yet been tested in thescanning research; and (4) our sample size is notoverly large; it is appropriate to use a less conser-vative criterion for statistical significance (Sauleyand Bedeian, 1989; Skipper, Guenther, and Nass,1967). We therefore selected 0.1, a priori, as theappropriate level of significance for testing ourhypotheses.

One additional element of our data analysisplan, however, helped further to increase the sta-tistical power of the tests. We conducted thehierarchical regression procedure described aboveusing all of the dependent variables at one time,in a ‘repeated measures’ regression. This tech-nique has been advocated by several authors asparticularly useful for increasing statistical powerin research areas—such as leadership, group, andorganization-level research—where attainingpower through large sample sizes is problematic(e.g., Cohen and Cohen, 1983; Hollenbeck, Ilgen,



and Sego, 1994). The technique helps to increasepower in at least two ways that are applicable forthe current study. First, the use of the multipleperformance measures (the repeated measures) inone regression provides the same benefits that aMANOVA does for a series of ANOVAs—theseries of individual ANOVAs/regressions is ‘pro-tected’ from chance-based Type I errors, and thedependent variables together may show a consis-tent pattern that produces statistical significance inthe multivariate analysis when such significancemay not be achieved in the individual analyses,thereby reducing Type II errors. The power advan-tage of repeated-measures regression relative toMANOVA, however, includes a gain from the useof continuous measures as independent variables(Hollenbeck et al., 1994). Second, the repeatedmeasures (i.e., within organizations) aspect of themultiple performance measures is used to increasethe power to test the between-organization fac-tors. This is accomplished by identifying the withinvariance and removing it from the total varianceto determine the actual variance that could beexplained by the between-organizations factors.This latter variance is used as the denominator ofthe F -test. ‘Sensitivity is gained in this instanceby removing the within-team variance that wouldotherwise have been treated as error variancein a typical between subject design (Cohen andCohen, 1983)’ (Hollenbeck et al., 1994: 9). Thus,

in testing for all of the performance measures atonce the power of the test is increased substantiallyrelative to a sequence of single-dependent-variableregressions.

RESULTS

Descriptive statistics and correlations are presentedin Table 3. Cronbach’s alpha scale reliabilitiesare also shown. To test our multivariate contin-gency hypotheses involving appropriate externaland internal scanning emphases in environmentsof differential dynamism, we conducted the hierar-chical regression procedure described above usingall of the dependent variables at one time ina ‘repeated-measures’ regression. The results areshown in Tables 4 and 5 and are briefly describednext.

The summary table (Table 4) shows that Step1—containing only the control and main effectvariables—explained a portion of the variance inthe three performance measures together. Addingthe double interaction terms in Step 2 did notincrease explanatory power. The addition in Step3 of the four specified triple interactions, how-ever, greatly increased the ability to explain per-formance. Table 5 allows more detailed interpreta-tions. For example, for a given level of scanningfirms that paid more attention to the task envi-ronment were higher performers. There were no

Table 3. Descriptive statistics and correlations

Variable α Mean S.D. 1 2 3 4 5 6 7 8 9

1. Logsize NA 4.29 0.442. Scanning level 0.81 3.64 1.40 0.12

Scanning emphasis3. External: task

environment0.61 5.20 1.12 −0.03 0.20∗

4. External: generalenvironment

0.71 3.83 1.38 00 0.16† 0.26∗∗

5. Internal: innovation 0.65 4.32 1.31 −0.04 0.20∗ 0.57∗∗∗ 0.34∗∗∗

6. Internal: efficiency 0.81 5.93 1.23 0.01 0.02 0.33∗∗∗ 0.14 0.077. Environmental

dynamism0.79 3.49 1.44 0.00 0.18† 0.15 0.16† 0.29∗∗ −0.18†

Performance8. Accounting return 0.91 3.50 1.10 0.06 0.06 −0.03 −0.10 −0.07 0.00 −0.129. Sales growth NA 3.61 1.15 0.28∗∗ 0.09 0.34∗∗∗ 0.14 0.29∗∗ 0.03 0.07 0.46∗∗∗

10. Overallperformance

NA 3.77 1.05 0.18† 0.13 0.09 −0.13 0.01 0.04 −0.16 0.72∗∗∗ 0.57∗∗∗

Note: n = 105; α = Cronbach’s alpha (except variables 6 and 8, where inter-item correlations are reported). †p < 0.1; ∗p < 0.05;∗∗p < 0.01; ∗∗∗p < 0.001



Table 4. Summary table for between firm effects for sales growth, accounting return, and overall performancesimultaneously (repeated measures regression)

Step 1 Step 2 Step 3

Between Error Between Error Between Error

SS 31.80 246.32 32.12 236.88 102.50 210.23R2 (adjusted R2) 0.13 (0.07) 0.14 (−0.01) 0.49 (.37)F 2.2∗ 1.12 5.21∗∗

d.f. (numerator/denominator) 7/96 15/88 19/84Change in R2 N.A. 0.01 0.35F for change in R2 N.A. 0.13 14.41∗∗

d.f. (numerator/denominator) N.A. 8/88 4/84G-G epsilon 0.89 0.89 0.90

∗p < 0.05; ∗∗p < 0.001

Table 5. Tests of hypotheses for between-firm effects for sales growth, accounting return, and overall performancesimultaneously (repeated measures)

IV # Variable Step 1 Step 2 Step 3

SS DF F SS DF F SS DF F

1 Logsize 13.02 1 5.07∗

2 Scanning level 0.93 1 0.363 Scanning emphasis on task

environment (EXTTASK)8.29 1 3.23†

4 Scanning emphasis on generalenvironment (EXTGEN)

2.44 1 0.95

5 Scanning emphasis on innovation(INTINV)

0.96 1 0.38

6 Scanning emphasis on efficiency(INTEFF)

2.19 1 0.85

7 Perceived Environmental dynamism(ENVDYN)

3.97 1 1.55

8 EXTTASK ∗ ENVDYN 0.17 1 0.069 EXTGEN ∗ ENVDYN 0.12 1 0.0510 INTINV ∗ ENVDYN 1.13 1 0.4211 INTEFF ∗ ENVDYN 1.07 1 0.412 EXTTASK ∗ INTINV 0.48 1 0.1813 EXTTASK ∗ INTEFF 0.54 1 0.214 EXTGEN ∗ INTINV 1.89 1 0.715 EXTGEN ∗ INTEFF 0.01 1 0.0016 EXTTASK ∗ INTINV ∗ ENVDYN 9.47 1 3.79†17 EXTTASK ∗ INTEFF ∗ ENVDYN 21.67 1 8.66∗∗

18 EXTGEN ∗ INTINV ∗ ENVDYN 10.51 1 4.2∗

19 EXTGEN ∗ INTEFF ∗ ENVDYN 12.87 1 5.14∗

†p < 0.1; ∗p < 0.05; ∗∗p < 0.01

significant effects on performance due to doubleinteractions. But, each of the four triple inter-actions involving an internal scanning domain,an external scanning domain, and environmentaldynamism was significantly related to the threeperformance measures together. This suggests thathigh performance requires the relative scanning

emphases within internal and external domainsboth must be appropriate to the level of dynamismin the external environment.

We cannot, however, use the beta weights ob-tained from multiple performance measures in therepeated-measures regression to accurately specifythe triple interaction relationships and compare



Table 6. Full model regression of sales growth

Variable Regression coefficient S.E.

Identification Value

1 Logsize a1 0.89 0.232 Scanning level a2 0.00 0.083 Scanning emphasis on task environment

(EXTTASK = X1)b1 0.33 0.13

4 Scanning emphasis on general environment(EXTGEN = X2)

b2 −0.03 0.09

5 Scanning emphasis on internal innovation(INTINV = X3)

b3 0.22 0.10

6 Scanning emphasis on internal efficiency(INTEFF = X4)

b4 −0.10 0.12

7 Environmental dynamism (ENVDYN = X5) b5 −0.10 0.088 EXTTASK∗ENVDYN b6 0.15 0.119 EXTGEN∗ENVDYN b7 0.02 0.06

10 INTINV∗ENVDYN b8 −0.03 0.0811 INTEFF∗ENVDYN b9 0.13 0.0912 EXTTASK∗INTINV b10 −0.20 0.0913 EXTTASK∗INTEFF b11 −0.13 0.0814 EXTGEN∗INTINV b12 −0.05 0.0715 EXTGEN∗INTEFF b13 0.05 0.0616 EXTTASK∗INTINV∗ENVDYN b14 0.15∗ 0.0717 EXTTASK∗INTEFF∗ENVDYN b15 0.25∗∗∗ 0.0618 EXTGEN∗INTINV∗ENVDYN b16 −0.13∗∗ 0.0519 EXTGEN∗INTEFF∗ENVDYN b17 −0.21∗∗ 0.07

∗p < 0.05; ∗∗p < 0.01; ∗∗∗ p < 0.001Interpretable regression coefficients are in bold font.

them to those we hypothesized. To allow us toexplain and graph the triple interactions of interest,Table 6 shows the full regression model for justone dependent performance variable: sales growth.We provide a detailed explanation of how theinteraction found in our data compared to theassertions of Hypothesis 1. Then, we offer a similarbut briefer explanation of the interaction associatedwith Hypothesis 2.

The interaction of task environment scanningemphasis, perceived environmental dynamism,and innovation scanning emphasis

The partial derivative of the full regression modelover scanning emphasis on task environment (X1)

yields the following equation:

�Y/�X1 = b1 + b6X5 + b10X3 + b11X4

+ b14X3X5 + b15X4X5 (2a)

Or, in terms of the beta values obtained in our data(see Table 6):

�Y/�X1 = 0.33 − 0.15X5 − 0.20X3 − 0.13X4

+ 0.15X3X5 + 0.25X4X5 (3a)

The effect on sales growth (Y) of changing scan-ning emphasis on task environment (X1) is a func-tion of perceived environmental dynamism (X5)

and scanning emphasis on innovation (X3). Weexamined the relationship by setting perceivedenvironmental dynamism first at its highest valueand then at its lowest value in our data. We held thevalue of scanning emphasis on efficiency constantat its mean.

For the highest perceived environmental dyn-amism, Equation 3a reduces to:

�Y/�X1 = 0.75 + 0.23X3 (4a)

Equation 4 will yield zero when X3 has a valueof −3.23. Thus, the inflection point of the slopeis −3.23. Because the inflection point of the



slope is within the range of values we observedin our data (−3.32 to 2.35), we concluded thatscanning emphasis on task environment has a non-monotonic effect on sales growth over the rangeof scanning emphasis on innovation when per-ceived environmental dynamism is the highest.When scanning emphasis on innovation is above−3.23, the equation will be positive but whenscanning emphasis on innovation is below −3.23the equation will be negative. The nonmonotonicrelationship indicates that when perceived environ-mental dynamism is the highest, scanning empha-sis on task environment has a positive effect onsales growth for all values of scanning emphasison innovation above −3.23 and a negative effecton sales growth for all values of scanning empha-sis on innovation below −3.23. We plotted thisrelationship in Figure 1. To plot the graph, we sub-stituted the least and the highest values of scanningemphasis on innovation in our data in Equation 4a.The graph shows (see the square markers) that theinflection point is very close to the extreme lowvalue of scanning emphasis on innovation. There-fore, the relationship is positive over practically

the entire range of scanning emphasis on inno-vation in our data. But, because the slope is anincreasingly positive one, we concluded that whenperceived environmental dynamism is the highest,the positive impact of scanning emphasis on taskenvironment on sales growth increases as scanningemphasis on innovation increases. This result sup-ports our first hypothesis.

Schoonhoven (1981: 353) has pointed out that‘an assumption of symmetrical effects is hidden inthe language of contingency theory’ and has sug-gested investigating symmetrical effects when test-ing contingency hypotheses such as those at hand.In our case, a symmetrical effect would be demon-strated if at the lowest perceived environmentaldynamism the slope of sales growth on scanningemphasis on task environment is decreasingly pos-itive and/or increasingly negative. To examine ifthis is indeed the case in our data, we substitutedthe extreme low value of perceived environmentaldynamism in our data in Equation 3a, which thenreduces to:

�Y/�X1 = −0.05 − 0.59X3 (5a)

1.90

0.99

-0.02

0.50

-0.50

-1.50

-2.50

X Axis : Scanning Emphasis on Innovation

1.30

-0.14

-1.43

1.50

2.50

Y Axis : d(Sales Growth)/d(Scanning Emphasis onTask Environment)

-4.00 -2.00-3.00 -1.00 0.00 1.00 2.00 3.00 4.00

High Dynamism Medium DynamismLow Dynamism

Figure 1. The effect of change in scanning emphasis on innovation and change in environmental dynamism on therelationship between sales growth and scanning emphasis on task environment



Equation 5a will yield zero when X3 has a valueof −0.09. Thus, the inflection point of the slopeis −0.09. Because the inflection point of the slopeis very close to the mean value (0, after center-ing) of scanning emphasis on innovation in ourdata, we concluded that at the lowest perceivedenvironmental dynamism scanning emphasis ontask environment has a nonmonotonic effect onsales growth over the range of scanning empha-sis on innovation. When scanning emphasis oninnovation is below −0.09, the equation will bepositive; when scanning emphasis on innovationis above −0.09, the equation will be negative.The nonmonotonic relationship indicates that whenperceived environmental dynamism is the lowestscanning emphasis on task environment has a pos-itive effect on sales growth for all values of scan-ning emphasis on innovation below −0.09 and anegative effect on sales growth for all values ofscanning emphasis on innovation above −0.09.Again, we plotted this relationship in Figure 1 (seethe triangular markers). At all values of scanningemphasis on innovation below the inflection point,the slope is a decreasingly positive one. Further,at all values of scanning emphasis on innova-tion above the inflection point, the slope is anincreasingly negative one. We concluded that whenperceived environmental dynamism is the lowest,(a) the positive impact of scanning emphasis ontask environment on sales growth decreases asscanning emphasis on innovation increases up toits inflection value, and (b) the negative impact ofscanning emphasis on task environment on salesgrowth increases as scanning emphasis on inno-vation increases beyond its inflection value. Thissymmetrical result further supports our first con-tingency hypothesis.

By depicting the two graphs in the same figure,we have tried to clearly demonstrate the interac-tion proposed in Hypothesis 1. In order to fur-ther increase clarity, we additionally plotted therelationship at the mean value of perceived envi-ronmental dynamism (see the diamond markers).Although infinite graphs are possible (each in adifferent plane, although they may appear to inter-sect when presented in a two-dimensional space),just one additional graph would serve to establishthe direction of rotation of the slope as perceivedenvironmental dynamism changes as a continu-ous, rather than a dichotomous variable. Thus, wesee that for any unit increase in scanning empha-sis on innovation, the slope of sales growth on

scanning emphasis on task environment is eitherdecreasingly negative or increasingly positive asperceived environmental dynamism increases fromits lowest to the mean to the highest value. Thissupports our first hypothesis. There is performancegain to increasing scanning emphasis on task envi-ronment when scanning emphasis on innovationand perceived environmental dynamism increasesimultaneously.

The interaction of general environmentscanning emphasis, perceived environmentaldynamism, and efficiency scanning emphasis

Analogous to the analysis of Hypothesis 1, thepartial derivative of the full regression model overscanning emphasis on general environment (X2)

yields the following equation:

�Y/�X2 = b2 + b7X5 + b12X3 + b13X4

+ b16X3X5 + b17X4X5 (2b)

Or, in terms of the beta values obtained in ourdata (see Table 6):

�Y/�X2 = −0.03 + 0.02X5 − 0.05X3 + 0.05X4

− 0.13X3X5 − 0.21X4X5 (3b)

For the highest perceived environmental dy-namism, Equation 3b reduces to:

�Y/�X2 = 0.03 − 0.41X4 (4b)

For the lowest perceived environmental dy-namism, Equation 3b reduces to:

�Y/�X2 = −0.08 + 0.28X4 (5b)

Equations 4b and 5b have been graphed inFigure 2 (see the square and triangular markers,respectively), along with an additional graph forthe relationship at the mean value of perceivedenvironmental dynamism (see the diamond mark-ers). Figure 2 depicts the following results:

• Scanning emphasis on the general environmenthas a nonmonotonic effect on sales growth overthe range of scanning emphasis on efficiency atall values of perceived environmental dynamismin our data.



2.06

0.22

-1.45

0.50

-1.50

-2.50

X Axis : Scanning Emphasis on Efficiency

0.22

-0.42

-0.08

-0.50

1.50

2.50

Y Axis : d(Sales Growth)/d(Scanning Emphasis on GeneralEnvironment)

-4.00-5.00 -2.00-3.00 -1.00 0.00 1.00 2.00 4.003.00 5.00

High Dynamism Medium DynamismLow Dynamism

Figure 2. The effect of change in scanning emphasis on efficiency and change in environmental dynamism on therelationship between sales growth and scanning emphasis on general environment

• When perceived environmental dynamism is thelowest (see the triangular markers), the slopeof sales growth on scanning emphasis on gen-eral environment is either decreasingly negative(below the inflection point of scanning empha-sis on efficiency) or increasingly positive (abovethe inflection point of scanning emphasis onefficiency). We concluded that when perceivedenvironmental dynamism is the lowest, then(a) the negative impact of scanning emphasis ongeneral environment on sales growth decreasesas scanning emphasis on efficiency increasesup to its inflection value, and (b) the posi-tive impact of scanning emphasis on generalenvironment on sales growth increases as scan-ning emphasis on efficiency increases beyond itsinflection value. This result supports our secondhypothesis.

• When perceived environmental dynamism is thehighest (see the square markers), the slope ofsales growth on scanning emphasis on gen-eral environment is either decreasingly positive(below the inflection point of scanning emphasison efficiency) or increasingly negative (above

the inflection point of scanning emphasis onefficiency). We concluded that when perceivedenvironmental dynamism is the highest, then(a) the positive impact of scanning emphasis ongeneral environment on sales growth decreasesas scanning emphasis on efficiency increasesup to its inflection value, and (b) the negativeimpact of scanning emphasis on general envi-ronment on sales growth increases as scan-ning emphasis on efficiency increases beyondits inflection value. This symmetrical result pro-vides further support to our second contingencyhypothesis.

• Finally, we look at the change in perceived envi-ronmental dynamism. From the rotation of thethree graphs, we observe that as perceived envi-ronmental dynamism decreases from its high-est to the mean to the lowest value, the slopeof sales growth on scanning emphasis on gen-eral environment is either decreasingly neg-ative or increasingly positive when scanningemphasis on efficiency increases. This againsupports our second hypothesis. There is per-formance gain to increasing scanning emphasis



on general environment when scanning empha-sis on efficiency increases as perceived environ-mental dynamism simultaneously decreases.

DISCUSSION

We used the concepts of dominant logic and sec-tor salience to develop predictions about whichinternal capabilities and which sectors of the exter-nal environment should receive relatively more orless CEO scanning emphasis in competitive envi-ronments that are overall more stable or moredynamic. We found that simultaneous scanningemphases on those internal capabilities and onthose external environment sectors appropriate tothe level of dynamism in the external environ-ment produced higher performance. Higher salesgrowth, for example, resulted for the manufactur-ing firms in our sample when CEOs facing moredynamic competitive environments simultaneouslyincreased their relative scanning emphases on thetask sectors of the external environment and oninnovation functions in the internal environment.Higher sales growth also occurred when CEOsfacing more stable environments simultaneouslyincreased their relative scanning emphases on thegeneral sectors of the external environment andon efficiency functions in the internal environ-ment. These results were not as strong, however,for return or overall performance. It may be thatthe benefits obtained from acute knowledge of theappropriate fit among key external and internalenvironment sectors pertain more to transactionsacross firm boundaries (i.e., the top line), and lessto profitability (i.e., the bottom line).

Overall, the hypotheses received strong support.Both of the predicted triple interactions were sig-nificantly related to sales growth, profitability, andoverall performance, together. And, the two tripleinteractions we graphed for sales growth alonewere significant in the precise directions hypothe-sized. This level of support is far greater than thatwhich would be expected merely by chance.

Our results have important implications forresearch. First, future studies should include boththe external and internal environments of thefirm when evaluating CEO scanning. Our studyfound that a simultaneous match among rela-tive scanning emphasis on external sectors, rel-ative scanning emphasis on internal sectors, and

dynamism in the external environment was associ-ated with firm performance. Research that empha-sizes only one or two of these aspects, as hasthe earlier scanning work, risks obtaining incom-plete or misleading findings. Second, these resultson CEO scanning support multivariate contin-gency and configural theories of strategy (e.g.,Miller and Mintzberg, 1984). Configurations mayexert their influence from the earliest stages ofthe strategy-making process. Future studies likelyshould develop and test more complex contin-gency and configural theories of CEO scanning.Such theories might explain further and test whya simultaneous match among scanning emphasison external sectors, scanning emphasis on internalsectors, and dynamism in the external environmentmay be more strongly associated with sales growththan profitability. Such research may profitablyinclude variables affecting transactions within firmboundaries—e.g., decision-specific and manage-rial factors (Rajagopalan et al., 1997), and struc-tural factors (Ocasio, 1997; Wally and Baum,1994)—when developing and testing new contin-gency theories of scanning and strategy-makingprocesses.

This study also has implications for managers.First, the ‘scanning selection’ task is importantto firm performance. Scanning emphasis on par-ticular sectors of the environment is necessaryfor an effective strategy-making process. Second,the specific sectors on which to focus depend onthe level of dynamism in the firm’s competitiveenvironment. Thus, top managers of firms facingdynamic environments should emphasize scanningthe external task sectors and the internal sectorsassociated with innovation. Top managers of firmsfacing relatively stable environments, however,should emphasize scanning external general sec-tors and internal sectors associated with efficiency.

Our study has a number of limitations. Sinceour data were gathered from the CEO—a sin-gle respondent—in each organization, the likeli-hood of common response bias must be evaluated.Doty, Glick, and Huber have noted that ‘Com-mon methods variance creates a problem when-ever the same informants (or method) providedata for both the independent and the dependentvariables in a study and a pattern of responseson the independent variables obviously and logi-cally implies a pattern of responses on the depen-dent variable’ (Doty, Glick, and Huber, 1993:1240, emphasis in original). They went on to



argue that, since the fit operationalizations andconfigural hypotheses in their study were quitecomplex, it seemed unlikely that the respondents‘could structure their responses’ (Doty et al., 1993:1240) to produce the hypothesized results. Giventhe complexity of the variable operationalizations(see below) and the relationships hypothesizedin our study—wherein the independent variableswere formed from factor analyses of respon-dents’ ratings, and the hypotheses involved rela-tionships moderated by two other variables simul-taneously—it appears similarly unlikely that meth-ods bias could be the source of results supportingthe hypotheses.

The sample was restricted to smaller manufac-turing firms located in a major southwestern state.Although the threat to external validity arisingfrom geographical restriction is likely quite low,our results may not be generalizable to either largemanufacturers or to service firms. Also, samplesize was limited due to the restricted researchbudget and the strict criteria for firm selection.This limitation may have relatively little influenceon confidence in our results, however, when oneconsiders that some recent scanning studies havehad much smaller sample sizes, that we used therelatively powerful repeated measures regressiontechnique (Cohen and Cohen, 1983; Hollenbecket al., 1994), and that it is generally quite diffi-cult to detect moderating effects in field studies(McClelland and Judd, 1993). In addition, our dataare cross-sectional, thereby making even strongcausal arguments more tentative. Conservatively,the results could be interpreted to provide prelim-inary evidence for links between configural exec-utive scanning behavior and firm performance.

Lastly, all the data we analyzed reflect the sub-jective estimates of manufacturing firm CEOs:first, because our interests centered on CEO per-ceptions and resulting outcomes; and second, be-cause the strict controls for allowing firms into thesample population effectively eliminated diversi-fied, publicly held firms with their published finan-cial data. Common methods variance is always apotential threat to validity in such situations. Thethreat posed to our results due to possible commonmethods variance is limited, however, by the factthat the hypotheses involved triple interactions.In order to ‘second-guess’ the hypotheses, sub-jects would have been required to purposely matchtheir responses to the environmental dynamism

scale with the importance they devoted to partic-ular external and internal environmental segmentsand to their responses to performance questions,without any cues as to the predicted ‘appropriate’matches required for superior performance. Giventhe complexity of the hypotheses, it is unlikelythat CEOs could have somehow ‘structured’ theirresponses to performance questions to reflect pre-vious responses to the multiple items that measuredthe predictor variables (e.g., Doty et al., 1993).

CONCLUSION

Top managers cannot take effective action to influ-ence organizational processes and outcomes untilthey form appropriate judgments: (1) about thelevels of key variables inside and outside theirfirm; and (2) about the causal relationships of thesevariables with one another and with firm perfor-mance (Priem and Harrison, 1994). In order todevelop these judgments, managers first must pri-oritize those external and internal areas deservingtheir attention, and then gather and interpret themost critical data. Scanning is therefore the firststep in the chain of activities leading to organiza-tional adaptation (Hambrick, 1981).

Our study suggests that: (1) executive scanningof appropriate sectors in both the external andinternal environments is important to firm perfor-mance; and (2) the pertinence of a particular sectordepends in part upon the overall level of dynamismin the external environment facing the firm. Exec-utives who do not scan both internally and exter-nally, or who do not appropriately prioritize envi-ronmental sectors, will likely be hampered in form-ing effective judgments about their firm’s competi-tive situation. Their subsequent actions could causefirm performance to suffer.

ACKNOWLEDGEMENTS

We thank John Butler, Greg Dess, Dave Harrison,Abdul Rasheed, Olivia Wang Lihua, and GeoffWaring for beneficial comments on prior drafts ofthis paper.

REFERENCES

Aguilar FJ. 1967. Scanning the Business Environment .Macmillan: New York.



Andrews KR. 1980. The Concept of Corporate Strategy .Irwin: Homewood, IL.

Ansoff IH. 1965. Corporate Strategy . McGraw-Hill: NewYork.

Ansoff IH. 1975. Managing strategic surprise by responseto weak signals. California Management Review18(2): 21–33.

Barney J. 1991. Firm resources and sustained competitiveadvantage. Journal of Management 17: 99–120.

Bluedorn AC, Johnson RA, Cartwright DK, BarringerBR. 1994. The interface and convergence of thestrategic management and organizational environmentdomains. Journal of Management 20: 201–262.

Bobko P. 1995. Correlation and Regression: Principlesand Applications for Industrial/Organizational Psy-chology and Management . McGraw-Hill: New York.

Bourgeois LJ. 1980. Strategy and environment: aconceptual integration. Academy of ManagementReview 5: 25–39.

Bourgeois LJ. 1985. Strategic goals, perceived uncer-tainty, and economic performance in volatile environ-ments. Academy of Management Review 10: 548–573.

Boyd KB, Fulk J. 1996. Executive scanning andperceived uncertainty: a multidimensional model.Journal of Management 22: 1–21.

Boyd KB, Dess GG, Rasheed AMA. 1993. Divergencebetween perceptual and archival measures of theenvironment: causes and consequences. Academy ofManagement Review 18: 204–226.

Burns T, Stalker GM. 1961. The Management ofInnovation . Tavistock: London.

Cameron K. 1978. Measuring organizational effective-ness in institutions of higher education. AdministrativeScience Quarterly 23: 604–632.

Chakravarthy B. 1986. Measuring strategic performance.Strategic Management Journal 7(5): 437–458.

Child J. 1972. Organization structure, environment, andperformance: the role of strategic choice. Sociology 6:1–22.

Child J. 1997. Strategic choice and the analysis of action,structure, organizations and environment: retrospectand prospect. Organizational Studies 18(1): 43–76.

Cohen J, Cohen P. 1983. Applied Multiple Regres-sion/Correlation Analyses for the Behavioral Sciences .Erlbaum: Hillsdale, NJ.

Cyert RM, March JG. 1963. A Behavioral Theory of theFirm . Prentice-Hall: Englewood Cliffs, NJ.

Daft RL, Sormunen J, Parks D. 1988. Chief executivescanning, environmental characteristics, and companyperformance: an empirical study. Strategic Manage-ment Journal 9(2): 123–139.

Darlington R. 1990. Regression and Linear Models .McGraw-Hill: New York.

Dess GG, Beard DW. 1984. Dimensions of organi-zational task environments. Administrative ScienceQuarterly 29: 52–73.

Dess GG, Ireland RD, Hitt MA. 1990. Industry effectsand strategic management research. Journal ofManagement 16: 5–25.

Dess GG, Robinson RB. 1984. Measuring organizationperformance in the absence of objective measures:the case of the privately-held firm and conglomerate

business unit. Strategic Management Journal 5(3):265–273.

Dill WR. 1958. Environment as an influence on man-agerial autonomy. Administrative Science Quarterly 3:409–443.

Doty DH, Glick WH, Huber GP. 1993. Fit, equifinality,and organizational effectiveness: a test of twoconfigurational theories. Academy of ManagementJournal 36: 1196–1250.

Duncan RB. 1972. Characteristics of organizational envi-ronments and perceived environmental uncertainty.Administrative Science Quarterly 17: 313–327.

Dutton JM, Freedman RD. 1984. Uncertainty and firms’strategic behaviors. Working paper, New YorkUniversity.

Eisenhardt KM. 1989. Making fast strategic decisions inhigh-velocity environments. Academy of ManagementJournal 12: 543–576.

Eisenhardt KM, Martin JA. 2000. Dynamic capabilities:what are they? Strategic Management Journal , SpecialIssue 21(10–11): 1105–1121.

El Sawy OA. 1985. Personal information systems forstrategic scanning in turbulent environments: can theCEO go on-line? MIS Quarterly 9: 53–60.

Fredrickson JW, Mitchell TR. 1984. Strategic decisionprocesses: comprehensiveness and performance in anindustry with an unstable environment. Academy ofManagement Journal 27: 399–423.

Glick WH, Miller CC, Huber GP. 1993. The impactof upper echelon diversity on organizational perfor-mance. In Organizational Change and Redesign: Ideasand Insights for Improving Performance, Huber GP,Glick WH (eds). Oxford University Press: New York;176–214.

Goodstein LD, Pfeiffer JW, Nolan TM. 1991. Appliedstrategic planning: a revised model for organizationalgrowth and vitality. In Strategic Planning: SelectedReadings , Pfeiffer JW (ed). Pfeiffer: San Diego, CA;ix–xxxix.

Hambrick DC. 1981. Specialization of environmentalscanning activities among upper level executives.Journal of Management Studies 18: 299–320.

Harrigan KR. 1983. Research methodologies for contin-gency approaches to business strategy. Academy ofManagement Review 8: 398–405.

Hatten KJ, Schendel DE, Cooper AC. 1978. A strategicmodel of the U.S. brewing industry. Academy ofManagement Journal 21: 592–610.

Hofer CW, Schendel D. 1978. Strategy Formulation:Analytical Concepts . West: St Paul, MN.

Hollenbeck JR, Ilgen DR, Sego DJ. 1994. Repeatedmeasures regression: enhancing the power ofleadership research. Leadership Quarterly 5: 3–23.

Jaccard J, Turrisi R, Wan CK. 1990. Interaction Effectsin Multiple Regression . Sage: Newbury Park, CA.

Kessler EH, Chakrabarti AK. 1996. Innovation speed:a conceptual model of context, antecedents, andoutcomes. Academy of Management Review 21:1143–1191.

Kim J, Mueller CW. 1978. Factor Analysis: StatisticalMethods and Practical Issues . Sage: Beverly Hills,CA.



Kumar N, Stern LW, Anderson JW. 1993. Conductinginterorganizational research using key informants.Academy of Management Journal 36: 1633–1651.

Learned EP, Christensen CR, Andrews KR, Guth WP.1965. Business Policy: Text and Cases . Irwin:Homewood, IL.

Lippman SA, Rumelt R. 1982. Uncertain imitability: ananalysis of interfirm differences in efficiency undercompetition. Bell Journal of Economics 13: 418–438.

Loehlin JC. 1992. Latent Variable Models . Earlbaum:Hillsdale, NJ.

March JG, Simon HA. 1958. Organizations . Wiley: NewYork.

McClelland GH, Judd CM. 1993. Statistical difficultiesof detecting interactions and moderator effects.Psychology Bulletin 114: 376–390.

Miles R, Snow C. 1978. Organizational Strategy, Struc-ture and Process . McGraw-Hill: New York.

Miller D. 1987. Strategy making and structure: analysisand implications for performance. Academy ofManagement Journal 30: 7–32.

Miller D. 1988. Relating Porter’s business strategies toenvironment and structure: analysis and performanceimplications. Academy of Management Journal 31:280–308.

Miller D, Droge C. 1986. Psychological and traditionaldeterminants of structure. Administrative ScienceQuarterly 31: 539–560.

Miller D, Friesen PH. 1984. Organizations: A QuantumView . Prentice-Hall: Englewood Cliffs, NJ.

Miller D, Mintzberg H. 1984. The case for configuration.In Beyond Method , Morgan G (ed). Sage: BeverlyHills, CA; 57–73.

Miller D, Shamsie J. 1996. The resource-based view ofthe firm in two environments: the Hollywood filmstudios from 1936 to 1965. Academy of ManagementJournal 39: 519–543.

Miller D, Toulouse JM. 1986. Chief executive personalityand corporate strategy and structure in small firms.Management Science 32: 1389–1409.

Milliken F. 1987. Three types of perceived uncertaintyabout the environment: state, effect, and responseuncertainty. Academy of Management Review 12:133–143.

Ocasio W. 1997. Towards an attention-based view of thefirm. Strategic Management Journal , Summer SpecialIssue 18: 187–206.

Pfeffer J, Salancik G. 1978. The External Control ofOrganizations: A Resource Dependence Perspective.Harper & Row: New York.

Porter ME. 1980. Competitive Strategy . Free Press: NewYork.

Porter ME. 1985. Competitive Advantage. Free Press:New York.

Prahalad CK, Bettis RA. 1986. The dominant logic:a new linkage between diversity and performance.Strategic Management Journal 7(6): 485–501.

Priem RL, Cycyota CS. 2001. On strategic judgment.In Handbook of Strategic Management , Hitt M,Freeman E, Harrison J (eds). Blackwell: London;493–519.

Priem RL, Harrison DH. 1994. Exploring strategic judg-ment: methods for testing the assumptions of pre-scriptive contingency theories. Strategic ManagementJournal 15(4): 311–324.

Priem RL, Rasheed AMA, Kotulic AG. 1995. Rational-ity in strategic decision processes, environmentaldynamism and firm performance. Journal of Manage-ment 21: 913–929.

Rajagopalan N, Rasheed AMA, Datta DK. 1993. Strate-gic decision processes: critical review and future direc-tions. Journal of Management 19: 349–384.

Rajagopalan N, Rasheed AMA, Datta DK, Spreitzer GM.1997. An integrative model of strategic decision mak-ing processes. In Strategic Decisions: Context, Pro-cess, and Outcomes , Papadakis V, Barwise P (eds).Kluwer: Boston, MA; 229–249.

Ratneshwar S, Warlop L, Mick DG, Seeger G. 1997.Benefit salience and consumers’ selective attention toproduct features. International Journal of Research inMarketing 14: 245–259.

Robinson RB, Pearce JA. 1988. Planned patterns ofstrategic behavior and their relationship to businessunit performance. Strategic Management Journal 9(1):43–60.

Rumelt RP. 1974. Strategy, Structure, and EconomicPerformance. Harvard Graduate School of BusinessAdministration: Cambridge, MA.

Sauley KS, Bedeian AG. 1989. .05: a case of the tailwagging the distribution. Journal of Management 15:335–344.

Schoonhoven CB. 1981. Problems with contingencytheory: testing assumptions hidden within thelanguage of contingency theory. AdministrativeScience Quarterly 26: 349–377.

Seidler J. 1974. On using informants: a techniquefor collecting quantitative data and controllingfor measurement error in organizational analysis.American Sociological Review 39: 816–831.

Simon HA. 1957. Administrative Behavior (2nd edn).Macmillan: New York.

Skipper JK Jr, Guenther AL, Nass G. 1967. The sacred-ness of .05: a note concerning the uses of statisticallevels of significance in social science. American Soci-ologist 2: 16–18.

Sutcliffe KM. 1994. What executives notice: accurateperceptions in top management teams. Academy ofManagement Journal 37: 1360–1378.

Thomas JB, Clark SM, Gioia DA. 1993. Strategicsensemaking and organizational performance: linkagesamong scanning, interpretation, action, and outcomes.Academy of Management Journal 36: 239–270.

Venkatraman N, Ramanujam V. 1987. Measurement ofbusiness economic performance: an examination ofmethod convergence. Journal of Management 13:109–122.

Wally S, Baum JR. 1994. Personal and structuraldeterminants of the pace of strategic decision making.Academy of Management Journal 37: 932–956.

Wernerfelt B. 1984. A resource-based view of the firm.Strategic Management Journal 5(2): 171–180.

Zaltman G, Duncan R, Holbek J. 1973. Innovations andOrganizations . Wiley: New York.


Documents

Chief executive scanning emphases, environmental dynamism, and manufacturing firm performance