argote_ mcevily_reagans_2003_managing knowledge in organizations an integrative framework and review...

Managing Knowledge in Organizations:An Integrative Framework and Review of

Emerging Themes

Linda Argote • Bill McEvily • Ray ReagansGraduate School of Industrial Administration, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213Graduate School of Industrial Administration, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213

Columbia Business School, Columbia University, New York, New York [email protected] • [email protected] • [email protected]

In this concluding article to the Management Science special issue on “Managing Knowl-edge in Organizations: Creating, Retaining, and Transferring Knowledge,” we provide an

integrative framework for organizing the literature on knowledge management. The frame-work has two dimensions. The knowledge management outcomes of knowledge creation,retention, and transfer are represented along one dimension. Properties of the context withinwhich knowledge management occurs are represented on the other dimension. These proper-ties, which affect knowledge management outcomes, can be organized according to whetherthey are properties of a unit (e.g., individual, group, organization) involved in knowledgemanagement, properties of relationships between units or properties of the knowledge itself.The framework is used to identify where research findings about knowledge managementconverge and where gaps in our understanding exist. The article discusses mechanisms ofknowledge management and how those mechanisms affect a unit’s ability to create, retainand transfer knowledge. Emerging themes in the literature on knowledge management areidentified. Directions for future research are suggested.(Knowledge Management; Organizational Learning; Knowledge Transfer; Innovation; OrganizationalMemory )

IntroductionResearch on the topics of organizational learning andknowledge management has enjoyed an extended andprosperous history. The importance of these conceptsfor understanding the coordination of organizationalactivity can be traced back to the early writings of suchinfluential thinkers as Adam Smith, who described thepin-making example as an illustration of how special-ization promoted experience-based learning (Smith1776/1937); Alfred Marshall, who argued that goodideas are quickly picked up and discussed by oth-ers in regional agglomerations (Marshall 1920); andMax Weber, who described the ability of bureaucra-cies to learn from experience (Weber 1922/1978). Itwas the Carnegie school, best exemplified by the work

of Richard M. Cyert and James G. March (Cyert andMarch 1963), that transformed these rudimentary andlargely anecdotal observations into a formal theory oforganizational learning and knowledge management,however. Four decades later, the field is characterizedby a wealth of empirical evidence and a wide array oftheoretical perspectives.

The highly differentiated nature of organizationallearning and knowledge management is a hallmarkof the field and is evident in the multitude of dis-ciplinary perspectives brought to bear on the topic.As we noted in the introduction to this specialissue, research on organizational learning and knowl-edge management spans the disciplines of economics,information systems, organizational behavior and

0025-1909/03/4904/0571$05.001526-5501 electronic ISSN

Management Science © 2003 INFORMSVol. 49, No. 4, April 2003, pp. 571–582

Marius

Surligner

Marius

Surligner

Marius

Surligner

ARGOTE, McEVILY, AND REAGANSManaging Knowledge in Organizations

theory, psychology, strategic management, and soci-ology. This diversity has contributed to the rapidadvance of the field by cultivating the simultane-ous development of specialized areas of inquiry thatinvestigate different aspects of organizational learn-ing and knowledge management.

The heterogeneity of knowledge managementresearch raises important questions about the degreeof integration across disciplines and the extent towhich a truly cumulative body of knowledge isemerging. For instance, theoretical foundations oforganizational learning and knowledge managementrange from the psychological emphasis on cogni-tion to the focus of economics on market structureand competition to the sociological orientation towardsocial structure. As research continues to advancein each of the discipline-based subfields of orga-nizational learning and knowledge management, itbecomes increasingly important to consider the extentof integration across these separate traditions. With-out addressing the question of integration, we runthe risk of propagating a highly fractionated view oforganizational learning and knowledge management.Moreover, a limited appreciation of the links acrossdisciplinary perspectives can prove to be inefficient asresearchers fail to take advantage of ideas producedin other areas and simply “rediscover” what is knownalready.

In the remainder of this paper, we address the fol-lowing questions to assess the state of integrationof knowledge accumulated across the different disci-plines. Are there points of convergence in the field? Ifso, do we see stable and consistent findings from onediscipline that are replicated or reinforced by findingsfrom other disciplines? Are researchers from differentdisciplines investigating unrelated aspects of organi-zational learning and knowledge management or arethey treading the same ground? What are the cur-rent themes emerging from recent research? Given thesize and diversity of the literature, a comprehensivereview is beyond the scope of a single paper (for arecent review, see Argote 1999). Instead, we draw onthe work appearing in this special issue and otherrepresentative work in the field to map the terrainof research on organizational learning and knowl-edge management. Figure 1 provides such an orient-ing framework.

The figure represents a summary of our effortsto create a framework for organizing the literaturebased on the relative positioning of work alongtwo critical dimensions: knowledge management out-comes (knowledge creation, retention, and transfer)and properties of the knowledge management context(properties of units, properties of the relationshipsbetween units, and properties of knowledge). Theknowledge management outcomes are representedon the vertical axis of Figure 1. Knowledge creationoccurs when new knowledge is generated in orga-nizations. Knowledge retention involves embeddingknowledge in a repository so that it exhibits somepersistence over time. Knowledge transfer is evidentwhen experience acquired in one unit affects another.These outcomes are related. For example, for an orga-nization to transfer knowledge, the knowledge mustbe retained. Attempts to transfer knowledge can leadto the creation of new knowledge. For example, Song,Almeida, and Wu (2003) show how new knowledge,in the form of patents, is generated when knowl-edge transfers across organizations through personnelmovement.

Despite the diversity of research on knowledgemanagement, theoretical explanations can be orga-nized according to three properties of the contextwithin which knowledge management occurs: Prop-erties of units (e.g., an individual, a group, or anorganization), properties of the relationships betweenunits, and properties of the knowledge itself. The con-text dimension displayed along the horizontal axisof Figure 1 highlights the fact that different theo-ries of knowledge management give causal prior-ity to different contextual properties. For example,some researchers emphasize the properties of theunits themselves, such as their absorptive capacity(Cohen and Levinthal 1990). Other researchers, whilefully recognizing the importance of properties of theunits, emphasize properties of relationships betweenunits, such as the network structure in which theunits are embedded (Reagans and McEvily 2003). Stillother researchers emphasize the different propertiesof knowledge, such as tacitness (Nonaka 1991), thatpromote or inhibit knowledge management.

The framework in Figure 1 can be used to character-ize research in the knowledge management literature.

572 Management Science/Vol. 49, No. 4, April 2003

Marius

Surligner

Marius

Surligner

Marius

Surligner

Marius

Surligner


Figure 1 Theoretical Framework for Organizing Research on Organizational Learning and Knowledge Management

Creation

Retention

Transfer

Properties of

Units

Properties of

The Relationships

between Units

Properties of

Knowledge

Knowledge Management Context

Kn

ow

led

ge

Ma

na

gem

ent

Ou

tco

mes

Because many studies in the special issue analyzemore than one outcome and more than one contextualvariable, we do not place studies in the cells of Fig-ure 1. Rather we discuss relevant findings from eachstudy as we elaborate the framework. The frameworkcan be used to identify points of integration acrossdifferent traditions and gaps in our understanding ofknowledge management. We undertake this integra-tion in the next section.

Knowledge Management ContextProperties of UnitsMany explanations of effective knowledge manage-ment focus on properties of a particular unit. Theunit could be an organization, an individual insidethe organization, or a population of organizations.The key driver of effective knowledge management issome characteristic of the unit itself. For example, psy-chologists and sociologists who study knowledge cre-ation and transfer emphasize status as an importantproperty of units. Status can be a property of an indi-vidual, a firm, or even of a piece of intellectual prop-erty and is an important predictor of knowledge man-agement outcomes. Thomas-Hunt, Neale, and Odgen(2003) describe the importance of expert status in pre-dicting the kind of information that an individual

shares with a group. Similarly, Borgatti and Cross(2003) demonstrate the importance of expert statusin predicting knowledge transfer across individuals.Sine, Shane, and Di Gregorio (2003) also consider theimportance of social status and find that knowledgecreated by a high-status institution is more likely to belicensed than knowledge created by a low-status insti-tution, even when the institution’s past performancein licensing is taken into account. Taken together,these studies suggest that status is an important fac-tor that explains knowledge creation, retention andtransfer. Status is, of course, not the only property ofa unit that facilities knowledge management. Statusis highlighted because it illustrates a convergence offindings across disciplines.

Properties of Relationships Between UnitsAnother tradition gives priority to how units are con-nected to each other. This tradition is characterizedby two approaches. One approach focuses on thedyadic relationship between social units. That rela-tionship can vary along a key set of dimensions,including intensity of connection, communication orcontact frequency, and social similarity. Each dimen-sion of the relationship can impact the knowledgemanagement process. For example, two studies in this

Management Science/Vol. 49, No. 4, April 2003 573


special issue focus on the importance of direct rela-tionships for effective knowledge transfer. Uzzi andLancaster (2003) focus on the strength of connectionbetween a loan officer and an entrepreneur, whileSong, Almeida, and Wu (2003) focus on the similar-ity between scientists. The key outcome is knowledgetransfer and the predictor of successful transfer issome quality of the dyadic relationship.

A second approach emphasizes the pattern of con-nections between multiple units (Rulke and Galask-iewicz 2000). For instance, knowledge flow betweentwo individuals is eased when the individuals areembedded in a dense web of third-party connections(Reagans and McEvily 2003). Moreover, knowledge ismore likely to transfer between establishments thatare owned by the same parent organization or thatare affiliated through the same franchise or chain thanacross independent organizations (Darr et al. 1995,Baum and Ingram 1998). The structural effect is notlimited to interpersonal or ownership connections.The important relationships could be a transactivememory system or consensus about who knows what.Transactive memory systems (Wegner 1986) facilitateknowledge retention (Liang et al. 1995) and knowl-edge transfer (Borgatti and Cross 2003). Knowledgeretention and transfer is even more efficient whengroup members share a short-hand language (Weberand Camerer 2003).

Properties of KnowledgeKnowledge properties affect the rate at which knowl-edge is accumulated, how much of it is retained,where it is retained, and how easily it diffuses withinand across firm boundaries. Tacit knowledge, orknowledge that is difficult to articulate, is more chal-lenging to transfer than explicit knowledge (Nonaka1991), and is best transferred through rich communi-cation media such as observation rather than throughmore explicit media (Nadler et al. 2003). Similarly,knowledge that has not been codified is more diffi-cult to transfer than codified knowledge (Zander andKogut 1995). Knowledge that is not well-understoodor is high in “causal ambiguity” is also harder totransfer than less ambiguous knowledge (Szulanski1996).

Recent research on knowledge management under-scores the importance of several other dimensions ofknowledge. One dimension is whether knowledge isperceived as external or internal to the focal unit.Where boundaries are drawn affects the value placedon knowledge and its usefulness. Menon and Pfef-fer (2003) show that organizational members are morelikely to value knowledge from external, rather thaninternal, sources, perhaps because such a valuationelevates the members’ status. On the other hand,knowledge coming from units perceived to be part ofthe same organization is more likely to transfer andimprove the performance of a focal unit than knowl-edge coming from external sources (Darr et al. 1995,Kane et al. 2002).

The extent to which knowledge is shared by orga-nizational members or uniquely possessed by a mem-ber also affects its transfer. Previous research hasshown that knowledge that is uniquely possessed bya member is less likely to be mentioned, repeated, andattended to in group discussion than commonly heldknowledge (Stasser and Titus 1985). Thus, groupsoften fail to make full use of their informationalresources because they do not surface informationidiosyncratic to particular members. Thomas-Huntet al. (2003) show that this decision bias depends onthe expertise of a group member and the extent towhich the member is integrated into the group. Inter-estingly enough, social isolates with special exper-tise are more likely to share their unique knowl-edge than socially connected members with uniqueexpertise.

Another dimension of knowledge that recentresearch points to is whether knowledge is public orprivate (Uzzi and Lancaster 2003). Knowledge avail-able in the public domain through standard reportstends to be “hard” information. By contrast, privateknowledge, which is not equally available to all orguaranteed by third parties, is “soft” informationabout unpublished aspects of a firm. Uzzi and Lan-caster (2003) show that different types of relation-ships or ties are suited for transferring private versuspublic knowledge. Arm’s-length ties are better suitedfor transferring public knowledge, whereas embed-ded ties are more suitable for transferring privateknowledge.



The properties of the knowledge management con-text identified in Figure 1 summarize how researchershave addressed what influences knowledge manage-ment outcomes. But what influences an outcomeis different from why the outcome occurs. Variousmechanisms explain how and why a particular contex-tual feature influences learning and knowledge man-agement outcomes. Ability, motivation, and opportu-nity are three causal mechanisms used to explain whycertain contextual features affect knowledge manage-ment outcomes. We turn now to a discussion of thesecausal mechanisms.

Mechanisms of KnowledgeManagementJust as successful individual performance dependson an individual’s ability, motivation, and opportu-nities to perform, successful knowledge managementalso depends on ability, motivation, and opportu-nity. Properties of the knowledge management con-text could impact an individual’s ability to create,retain, or transfer knowledge. Or the context couldprovide people with the motives and incentives toparticipate in the knowledge management process.Or the context could provide an individual with theopportunity to create, retain, or transfer knowledge.Properties of the knowledge management context canoperate through more than one causal mechanism.For example, social relationships provide individualswith the opportunity to create, retain, and transferknowledge. Social relationships also provide individ-uals with the incentives to participate in the process.

AbilityAbility is an important part of the knowledge man-agement process. Abilities are innate but also canresult from training (Nadler et al. 2003). Training inanalogical reasoning, for example, increases an indi-vidual’s ability to transfer knowledge accumulated onone task to a related task (Gick and Holyoak 1983,Thompson et al. 2000). The similarity between thetasks makes that transfer easier (Darr and Kurtzberg2000). Experience also affects ability. Individuals andorganizational units have the capacity to understand

knowledge in areas where they have previous expe-rience because individuals learn, or absorb, knowl-edge by associating it with what they already know(Cohen and Levinthal 1990). Factors that increase aperson’s ability to manage knowledge need not bespecific to him or her. For example, transactive mem-ory systems (Wegner 1986) and common short-handlanguages (Weber and Camerer 2003) are propertiesof relationships that affect members’ ability to create,retain, or transfer knowledge.

MotivationRewards and incentives are important componentsof the knowledge management process. The “notinvented here” syndrome in organizations is an exam-ple of how rewards can affect knowledge manage-ment outcomes. Members of a unit are unlikely totransfer knowledge from other parts of the organi-zation if they are not rewarded for utilizing internalknowledge (Menon and Pfeffer 2003). Social rewardscan be just as important as monetary rewards. Strongties promote the transfer of tacit knowledge (Uzzi1997) because strong ties are more likely to be gov-erned by the norms of reciprocity. The cooperativenorms associated with social cohesion also facili-tate knowledge transfer (Reagans and McEvily 2003).Uncooperative behavior damages individuals’ reputa-tions, so they are willing to expend extra effort trans-ferring knowledge to protect their social standing.

OpportunityAbility and extra effort are even more valuable whencoupled with opportunity. Effective knowledge man-agement results from providing individuals with theopportunity to create, retain, and transfer knowledge.Those opportunities could result from direct or indi-rect experience. Experience provides individuals withthe opportunity to create knowledge through trial-and-error learning. Interruptions in experience pro-vide opportunities for knowledge transfer (Zellmer-Bruhn 2003).

Organizational relationships influence knowledgemanagement outcomes by providing members theopportunity to learn from each other. Organizationsreduce the amount of distance, either physically orpsychologically, between people. By reducing that



distance, organizations provide members with theopportunity to learn from each other. Learning byobservation is an example of such indirect learning.Instead of accumulating knowledge directly, an indi-vidual accumulates knowledge by watching anotherperson perform a task (Nadler et al. 2003). Proxim-ity also provides people with the opportunity to learnwho knows what, so members know where to searchfor relevant knowledge and information (Borgatti andCross 2003). The transfer of routines, tools, and tech-nology across units within organizations means thatmembers of a recipient unit benefit from knowledgeacquired at the first unit (Epple et al. 1996, Winter andSzulanski 2001).

Informal networks serve a similar function. Bymaking knowledge more proximate, informal tiespromote vicarious learning. Informal connectionsallow people to benefit from knowledge accumu-lated by close contacts and associates (Hansen 1999,McEvily and Zaheer 1999, Reagans and Zucker-man 2001, Uzzi and Lancaster 2003). Those contactscould be inside or outside the organization. Person-nel movement across organizations or organizationalunits also increases the opportunity for one unit tolearn from another (Almeida and Kogut 1999, Songet al. 2003).

Emergent ThemesThe framework in Figure 1 also highlights sev-eral important themes that are emerging on knowl-edge management in organizations. One theme isthe importance of social relations in understandingknowledge creation, retention, and transfer. A secondtheme is that knowledge management outcomes areaffected by the “fit” or congruence between proper-ties of knowledge, properties of units, and proper-ties of relationships between units. The significanceof where organizational boundaries are drawn forknowledge transfer is another theme. Work is alsoemerging on how different types of experience havedifferent effects on learning outcomes. The effect ofenvironmental factors on learning outcomes in firmsrepresents another theme. The importance of embed-ding organizational knowledge in a repository so thatit persists over time is a sixth theme. Each of these

themes is discussed and directions for future researchabout each theme are suggested.

Significance of Social RelationsRelative to research on how properties of units andproperties of knowledge affect outcomes, research onhow properties of relationship(s) between units affectlearning and knowledge management outcomes is anewer theme. As mentioned previously, research inthis area can be organized into two broad categories:Work focusing on the qualities of a given relationshipor dyad, and work emphasizing the properties of asocial system of relationships (i.e., networks or collec-tive entities).

Research characteristic of the dyadic approach hasbeen primarily directed at understanding how thecloseness or strength of a relationship between twoparties is related to the effectiveness of knowledgetransfer. For instance, Uzzi’s (1997) ethnographicstudy reveals that strong ties develop relationship-specific heuristics that ease the transfer of knowledge.Similarly, Hansen’s (1999) study of product develop-ment teams indicates that strong ties are conducive tothe transfer of complex knowledge, while weak tiesaid in the search for new knowledge.

Dyadic research on knowledge transfer could beusefully extended to also consider the effects of tiestrength on other properties of the knowledge man-agement context. For instance, in what ways mighttie strength affect the creation and retention of knowl-edge? Future research could also explore how thedegree of asymmetry among the members of a dyadaffects knowledge management outcomes. Researchby Krackhardt (1990) shows that the members ofa dyad frequently have differing perceptions of thestrength and intensity of their relationship. Howmight these inconsistencies influence the effectivenessof search, transfer, and related activities?

Beyond tie strength, dyadic research mightalso consider which other dimensions of relation-ships affect knowledge management outcomes. Forinstance, McEvily et al. (2003) argue that the levelof trust affects the extent of knowledge disclosure,screening, and sharing between two parties. Fromthis perspective, trust alleviates concerns aboutknowledge appropriation and misuse. Trust also



reduces apprehension about the veracity of knowl-edge, thereby diminishing the tendency to questionthe knowledge’s accuracy. An interesting complementto this line of work would be to explore some of the“dysfunctional” or negative aspects of relationships.Similarly, dyadic research could also investigateissues such as the effect of power or conflict onknowledge management outcomes.

Whereas dyadic research on knowledge manage-ment focuses on the properties of a single relationshipbetween two individuals or units, other research con-centrates on the properties of a set of relationships ina social system. One important line of work emergingin this area is on the structural configuration, or net-work, of a set of relationships. Properties of an orga-nization’s internal social network (Borgatti and Cross2003, Thomas-Hunt et al. 2003) as well as its networkto other firms (Uzzi and Lancaster 2003) affect learn-ing and knowledge transfer.

Although existing research points to the importantrole of informal networks for effective knowledgetransfer, the available empirical evidence on the asso-ciation between informal social networks and knowl-edge transfer is limited (see Reagans and McEvily2003). More research is needed on how informal net-works affect the knowledge management process. Forinstance, are some network configurations more effec-tive at creating and retaining knowledge than others?Moreover, do certain network positions endow theoccupants of those positions with differential advan-tages or liabilities relative to other positions in thenetwork?

In addition to social networks, other properties ofsocial systems have been associated with knowledgemanagement outcomes. For instance, different fea-tures of the formal organizational structure, such asthe extent to which a firm is integrated (Sorenson2003) or centralized (Chang and Harrington 2003),are associated with knowledge transfer. Culture isanother important social factor. Cultural factors, suchas a firm’s idiosyncratic conventions and specialized“homemade” language (Weber and Camerer 2003) orits learning orientation and values (Bunderson andSutcliff, forthcoming, Edmondson 1999), affect knowl-edge and performance outcomes.

“Fit” of Properties of ContextsThe concept of “fit” or congruence has recurred inmany theories in the social and organizational sci-ences (Burton et al. 2002), including theories of knowl-edge management (Birkinshaw et al. 2002). Person-jobfit predicts outcomes in the stress literature (Edwards1996). The fit between the organization and charac-teristics of its task or environment predicts organi-zational performance in structural contingency the-ory (Comstock and Scott 1977) and survival in pop-ulation ecology theory (Hannan and Freeman 1989).Consistent with these perspectives, the emerging liter-ature on knowledge management illustrates that thefit between properties of knowledge, units, relation-ships, and the environment predicts knowledge man-agement outcomes. For example, Das (2003) demon-strates that the fit between characteristics of a problemconfronting an organization and its problem-solving“moves” or approaches predicts performance. Uzziand Lancaster (2003) show that the fit between thenature of knowledge and the type of tie used totransfer it affects learning outcomes. Sorenson (2003)shows that the fit between the turbulence of the envi-ronment and the design of the firm predicts its prob-ability of survival.

These studies find better performance outcomeswhen components fit or are congruent with eachother. More research is needed on the mechanismthrough which fit affects learning and knowledgemanagement outcomes. For example, does fit affectopportunities to transfer knowledge by making mem-bers more aware of other knowledge from which theywould benefit? Or does fit affect ability by making theknowledge easier to understand? In addition, researchis needed to identify dimensions of fit and to specifya priori when components fit each other and whenthey do not.

Research is also needed on whether there are con-ditions under which having dissimilar componentsthat do not fit each other would be more benefi-cial for learning outcomes. For example, heteroge-neous group where members have different back-grounds have been found to be more creative thanhomogeneous groups where members are similar toeach other (e.g., see Lant et al. 1992, Lapre and vanWassenhove 2001, Williams and O’Reilly 1998). In this



volume, Song et al. (2003) find that personnel mobil-ity was more likely to result in interfirm knowledgetransfer and the creation of new knowledge at thefirm when hired engineers had technological exper-tise distant from that of the hiring firm. Thus, havingand bridging components that do not fit or are dif-ferent from each other can facilitate the creation ofknowledge.

Organizational BoundariesCurrent work on knowledge management also illus-trates that where organizational boundaries aredrawn has important implications for knowledgetransfer and subsequent organizational performance.Research has shown that knowledge is more likely totransfer across units that are part of the same organi-zation or superordinate relationship (Darr et al. 1995,Baum and Ingram 1998, Ingram and Simons 2002).Similarly, Zellmer-Bruhn (2003) found that units aremore likely to transfer best practices from units thatare part of the same organization than from units thatbelong to a different organization. Song et al. (2003)present evidence that an organization is more likelyto benefit from the knowledge of individual scientistswhen those scientists move to the organization thanwhen they are employed by a different organization.Irwin and Klenow (1994) demonstrate that althoughknowledge transfers across firms in the semiconduc-tor industry, firms learn three times as much fromtheir own experience as from experience at anotherfirm. Thus, these studies suggest that organizationalunits are more likely to benefit from internal thanexternal knowledge.

In a somewhat different vein, Menon and Pfef-fer (2003) find that organizational members are morelikely to value knowledge from external than frominternal sources. The researchers suggest that indi-viduals might value external knowledge because theyare less familiar with its limitations and becausevaluing external knowledge would not reduce theirown status. The greater value organizational membersplace on external relative to internal knowledge, doesnot indicate that external knowledge is more likelyto improve organizational performance than inter-nal knowledge. Indeed, the cases Menon and Pfeffer

describe suggest that, by not valuing internal knowl-edge, the two organizations missed important oppor-tunities to improve their performance.

Being able to bridge knowledge boundaries is animportant capability (Reagans and McEvily 2003). Git-tleman and Kogut (2003) show that having scientistswho are able to bridge two very different knowledgebases—the logic of the scientific community and thelogic of patenting—increases firm innovation.

More research is needed on the conditions underwhich organizational members value internal versusexternal knowledge and the conditions under whichusing internal versus external knowledge is more (orless) likely to improve a unit’s performance. Furtherresearch is also needed on the mechanisms throughwhich organizational boundaries affect knowledgetransfer. For example, do boundaries affect member’ssocial identity which in turn affects knowledge trans-fer (e.g., Kane et al. 2002)? Do boundaries affect theextent to which knowledge is understood and thusaffect member’s ability to transfer knowledge? Doboundaries affect the rewards members receive andtheir motivation to transfer knowledge or do bound-aries affect member’s awareness of knowledge andthe opportunities to transfer it?

Nature of ExperienceUnderstanding the effects of different types of expe-rience on learning and knowledge management out-comes is a new direction identified in recent reviewsof the organizational learning literature (Argote andOphir 2002, Ingram 2002, Schulz 2002) that is repre-sented in this special issue. Researchers are taking amore fine-grained view of experience to identify thetypes of experience that facilitate, impede, or have noeffect on learning outcomes. For example, somewhatdifferent or diverse experience appears to be morebeneficial for learning than either identical experienceor very different experience (Schilling et al. 2003).

Determining how experience interacts with char-acteristics of the organization is an important cur-rent research trend. Whether organizations are spe-cialists or generalists affects their ability to learn fromexperience. Ingram and Baum (1997) show that spe-cialists that concentrate in a small number of geo-graphic areas are more likely to learn from their own



experience than generalists that operate over a largerarea. Similarly, Haunschild and Sullivan (forthcom-ing) report that specialists learn more from diverseexperience than generalists. Sorenson (2003) finds thatvertically integrated firms learn more from their expe-rience in turbulent environments than their noninte-grated counterparts.

Nadler et al. (2003) demonstrate that experienceobserving someone perform a task is more benefi-cial for subsequent performance than other types ofexperience, such as that acquired through classroomtraining. Experience individuals acquire by observingsomeone perform a task provides opportunities forthem to acquire tacit (Nonaka 1991) as well as explicitknowledge. Individuals who learn through observa-tion may not be able to articulate what they learn butare able to transfer the knowledge to a new task.

More research is needed on the mechanismsthrough which experience affects learning outcomes.For example, do different types of experience provideorganizational members with a better understandingof the task and thus, increase their ability to manageknowledge? Work is also needed on the conditionsunder which experience is most beneficial (or harm-ful) for learning outcomes. Research is also neededon how experience translates into the development ofcapabilities at firms (Eisenhardt and Martin 2000).

Environmental FactorsAnother theme of current knowledge managementresearch is that environmental factors affect learn-ing outcomes in firms. The turbulence in the envi-ronment (Sorenson 2003), the degree of competition(Chang and Harrington 2003) and the proportion ofcustomers with particular characteristics (Lee et al.2003) affect the success of learning strategies andorganizational designs. For example, Lee et al. (2003)find that the “exploration” of new technology is moreeffective than “exploitation” (March 1991) when theproportion of “power users” (i.e., users who are lessattentive to compatibility) is substantial or when anew technology is introduced before a more estab-lished one takes off. Sorenson (2003) finds that inte-grated firms benefit more from their experience thannonintegrated firms in turbulent contexts. Similarly,using a computational model, Chang and Harrington

(2003) find that centralized multiunit firms performbetter than decentralized ones when competition isintense.

Understanding “ecologies” of learning (Levitt andMarch 1988) or how learning by other firms (Ingram2002) or populations of firms (Miner and Haunschild1995) can affect a focal firm is an important issue thatwould benefit from future research. Understandingsuch interactions would have important implicationsfor the competitive behavior of firms.

Embedding KnowledgeAnother theme emerging from the literature is theembedding of knowledge in various reservoirs orrepositories. Knowledge can be embedded in individ-ual members, in the organization’s rules, routines, cul-tures, structures and technologies (Levitt and March1988, Starbuck 1992, McGrath and Argote 2001, Walshand Ungson 1991). Das (2003) describes how bothlocating and adapting solutions in the organization’smemory are important problem-solving moves usedto solve technical support problems. Drawing onwork about transactive memory systems or knowl-edge of who knows what, Borgatti and Cross (2003)show that knowledge of “who knows what” predictsinformation transfer in organizations.

More research is needed on how knowledge isembedded in an organization’s memory and the effectof where knowledge is embedded on performanceoutcomes. Understanding how transactive memorysystems develop and their consequences for groupperformance is an active research area that wouldbenefit from continued research (e.g., see Faraj andSproull 2000, Hollingshead 1998, Liang et al. 1995,Lewis forthcoming). Examining the process thoughwhich knowledge is embedded in rules and routinesand the effect of such embedding on group and orga-nizational outcomes is another important researcharea that would benefit from additional research(e.g., Cohen and Bacdayan 1994, March et al. 2000).More research is needed on how properties of units,properties of relationships, and properties of knowl-edge affect whether knowledge persists through timeor whether it depreciates (e.g., Benkard 2000, Darret al. 1995). Whether or not knowledge depreciateshas important implications for both operational andstrategic decisions of firms (Argote et al. 1990).



ConclusionThe papers in the special issue attest to the vitalityand diversity of research on organizational learningand knowledge management. Researchers from differ-ent disciplines, using different methods, and studyingdifferent contexts are increasing our understanding ofmanaging knowledge in organizations. The relianceon different disciplinary perspectives, different meth-ods, and different empirical contexts helps establishthe extent to which findings generalize and to identifythe boundary conditions under which they apply.

In spite of the diversity of disciplines, methodsand contexts, there is an impressive degree of inte-gration across research traditions. To organize theliterature and facilitate the identification of pointsof convergence, we proposed an integrative frame-work that highlights knowledge management out-comes and contextual properties as key dimensions.The framework identifies common areas of researchaccording to which contextual properties (of units,relationships, and knowledge) affect knowledge man-agement outcomes (creation, retention, and transfer).We also identify the key causal mechanisms (ability,motivation, and opportunity) that help explain howand why certain contextual properties affect knowl-edge management outcomes.

Stepping back from the framework in Figure 1, sev-eral research trends are evident. Across the rows ofthe figure, we can see that the majority of papers inthis issue focus on knowledge transfer. Thus, therewas more research in the bottom row of Figure 1 thanin the other two rows. Although there are certainlymore opportunities for research on knowledge trans-fer, knowledge creation, and retention warrant addi-tional attention as well.

Across the columns of Figure 1 we see that researchrepresented in the special issue is more balancedacross the columns than across the rows. Our sense,however, is that research on how properties of rela-tionships affect organizational learning and knowl-edge management is a more recent research trendthan research in the other columns. This is a verypromising development because relationships are crit-ical when one moves beyond studying individuals tostudying social units such as organizations.

The framework in Figure 1 also points to severalemerging themes that cut across different researchtraditions. Social relationships matter for knowledgecreation, retention, and transfer. When properties ofunits, properties of relationships and properties ofknowledge fit or are congruent with each other,knowledge retention, and transfer increase. Knowl-edge creation, by contrast, may be stimulated by alack of congruence or parts that do not fit together.Experience can be structured to promote learning out-comes in firms. Where boundaries are drawn mat-ters for knowledge creation, retention, and transfer.Features of the external environment affect learningoutcomes in firms. And embedding knowledge intransactive memory systems, short-hand languages,routines, technologies, and other knowledge reposito-ries can promote knowledge retention and transfer infirms.

The emerging themes that we have highlighted rep-resent exciting points of convergence and integrationacross disciplines. We hope that the framework devel-oped here is valuable in integrating the literature andsuggesting future research opportunities. We look for-ward to research that continues to develop the emerg-ing themes and further integrates work across dis-ciplines in order to advance the field of knowledgemanagement.

AcknowledgmentsThe authors wish to thank the National Science Foundation (GrantSES–0004283) for its generous support and Bruce Kogut and LeighThompson for their very helpful comments.

ReferencesAlmeida, P., B. Kogut. 1999. Localization of knowledge and the

mobility of engineers in regional networks. Management Sci. 45905–917.

Argote, L. 1999. Organizational Learning: Creating, Retaining andTransferring Knowledge. Kluwer, Norwell, MA., L., R. Ophir. 2002. Intraorganizational learning. J. A. C. Baum,ed. Companion to Organizations. Blackwell, Oxford, UK, 181–207., S. L. Beckman, D. Epple. 1990. The persistence and transferof learning in industrial settings. Management Sci. 36 140–154.

Baum, J. A. C., P. Ingram. 1998. Survival-enhancing learning inthe Manhattan hotel industry, 1898–1980. Management Sci. 44996–1016.


Marius

Surligner


Benkard, C. L. 2000. Learning and forgetting: The dynamics of air-craft production. Amer. Econom. Rev. 90 1034–1055.

Birkinshaw, J., R. Nobel, J. Ridderstrale. 2002. Knowledge as a con-tingent variable: Do the characteristics of knowledge predictthe organizational structure? Organ. Sci. 13 274–289.

Borgatti, S. P., R. Cross. 2003. A relational view of informationseeking and learning in social networks. Management Sci. 49(4)432–445.

Bunderson, J. S., K. M. Sutcliff. Management team orientation andbusiness unit performance. J. Appl. Psych. Forthcoming.

Burton, R. M., J. Lauridsen, B. Obel. 2002. Return on assets lossfrom situational and contingency misfits. Management Sci. 481461–1485.

Chang, M., J. E. Harrington, Jr. 2003. Multimarket competition, con-sumer search, and the organizational structure of multi-unitfirms. Management Sci. 49(4) 541–552.

Cohen, M. D., P. Bacdayan. 1994. Organizational routines are storedas procedural memory: Evidence from a laboratory study.Organ. Sci. 5 554–568.

Cohen, W. M., D. Levinthal. 1990. Absorptive capacity: A newperspective on learning and innovation. Admin. Sci. Quart. 35128–152.

Comstock, D. E., W. R. Scott. 1977. Technology and the structureof subunits: Distinguishing individual and workgroup effects.Admin. Sci. Quart. 22 177–202.

Cyert, R. M., J. G. March. 1963. A Behavioral Theory of the Firm.Prentice-Hall, Englewood Cliffs, NJ.

Darr, E., L. Argote, D. Epple. 1995. The acquisition, transfer anddepreciation of knowledge in service organizations: Productiv-ity in franchises. Management Sci. 41 1750–1762.

Darr, E. D., T. R. Kurtzberg. 2000. An investigation of partnersimilarity dimensions on knowledge transfer. Organ. BehaviorHuman Decision Processes 82 28–44.

Das, A. 2003. Knowledge and productivity in technical supportwork. Management Sci. 49(4) 416–431.

Edmondson, A. 1999. Psychological safety and learning behaviorin work teams. Admin. Sci. Quart. 44 350–383.

Edwards, J. R. 1996. An examination of competing versions ofthe person—environment fit approach to stress. Acad. Manage-ment J. 39 292–339.

Eisenhardt, K. M., J. A. Martin. 2000. Dynamic capabilities: Whatare they? Strategic Management J. 21 1105–1122.

Epple, D., L. Argote, K. Murphy. 1996. An empirical investigationof the micro structure of knowledge acquisition and transferthrough learning by doing. Oper. Res. 44 77–86.

Faraj, S., L. Sproull. 2000. Coordinating expertise in software devel-opment teams. Management Sci. 46 1554–1568.

Gick, M. L., K. J. Holyoak. 1983. Schema induction and analogicaltransfer. Cognitive Psych. 15 1–28.

Gittelman, M., B. Kogut. 2003. Does good science lead to valuableknowledge? Biotechnology firms and the evolutionary logic ofcitation patterns. Management Sci. 49(4) 366–382.

Hannan M. T., J. Freeman. 1989. Organizational Ecology. HarvardUniversity Press, Cambridge, MA.

Hansen, M. 1999. The search-transfer problem: The role of weakties in sharing knowledge across organization subunits. Admin.Sci. Quart. 44 82–111.

Haunschild, P. R., B. N. Sullivan. Learning from complexity: Effectsof accident, incident heterogeneity on airline learning. Admin.Sci. Quart. Forthcoming.

Hollingshead, A. B. 1998. Group and individual training: Theimpact of practice on performance. Small Group Res. 29254–280.

Ingram, P. 2002. Interorganizational learning. J. A. C. Baum, ed.Companion to Organizations. Blackwell, Oxford, U.K.

Ingram, P., J. A. C. Baum. 1997. Opportunity and constraint: Orga-nizations’ learning from the operating and competitive experi-ence of industries. Strategic Management J. 18 75–98.

Ingram, P., S. Simons. 2002. The transfer of experience in groups oforganizations: Implications for performance and competition.Management Sci. 48 1517–1533.

Irwin, D. A., P. J. Klenow. 1994. Learning by doing spillovers in thesemiconductor industry. J. Political Econom. 102 1200–1227.

Kane, A., L. Argote, J. L. Levine. 2002. Knowledge transfer betweengroups via personnel rotation: Effects of social identity andknowledge quality. Presented at the Academy of Managementmeetings, Denver, CO.

Krackhardt, D. 1990. Assessing the political landscape: Structure,cognition, and power in organizations. Admin. Sci. Quart. 35342–369.

Lant, T. K., F. J. Milliken, B. Batra. 1992. The role of manageriallearning and interpretation in strategic persistence and reori-entation: An empirical exploration. Strategic Management J. 13585–608.

M. A. Lapré, L. N. van Wassenhove. 2001. Creating and transferringknowledge for productivity improvement in factories. Manage-ment Sci. 47 1311–1325.

Lee, J., J. Lee, H. Lee. 2003. Exploration and exploitation in the pres-ence of network externalities. Management Sci. 49(4) 553–570.

Levitt, B., J. G. March. 1988. Organizational learning. Annual Rev.Sociology 14 319–340.

Lewis, K. Measuring transactive memory systems in the field: Scaledevelopment and validation. J. Appl. Psych. Forthcoming.

Liang, D., R. Moreland, L. Argote. 1995. Group versus individ-ual training and group performance: The mediating effects oftransactive memory. Personality Social Psych. Bull. 21 384–393.

March, J. G. 1991. Exploration and exploitation in organizationallearning. Organ. Sci. 2 71–87., M. Schulz, X. Xhoiu. 2000. The Dynamics of Rules: Studiesof Change in Written Organizational Codes. Stanford UniversityPress, Stanford, CA.

Marshall, A. 1920. Principles of Economics: An Introductory Volume.Macmillan, London, U.K.

McEvily, B., A. Zaheer. 1999. Bridging ties: A source of firm hetero-geneity in competitive capabilities. Strategic Management J. 201133–1156., V. Peronne, A. Zaheer. 2003. Trust as an organizing principle.Organ. Sci. 14 91–103.



McGrath, J. E., L. Argote. 2001. Group processes in organizationalcontexts. M. A. Hogg, R. S. Tindale, eds. Blackwell Handbookof Social Psychology, Vol. 3: Group Processes. Blackwell, Oxford,UK, 603–627.

Menon, T., J. Pfeffer. 2003. Valuing internal versus external knowl-edge. Management Sci. 49(4) 497–513.

Miner, A. S., P. R. Haunschild. 1995. Population-level learning. Res.Organ. Behavior 17 115–166.

Nadler, J., L. Thompson, L. Van Boven. 2003. Learning negotiationskills: Four models of knowledge creation and transfer. Man-agement Sci. 49(4) 529–540.

Nonaka, I. 1991. The knowledge-creating company. Harvard Bus.Rev. 69 96–104.

Reagans, R., B. McEvily. 2003. Network Structure and Knowl-edge Transfer: The Transfer Problem Revisited. Working paper,Columbia University, New York., E. W. Zuckerman. 2001. Networks, diversity and perfor-mance: The social capital of R&D units. Organ. Sci. 12 502–517.

Rulke, D. L., J. Galaskiewicz. 2000. Distribution of knowledge,group network structure, and group performance. ManagementSci. 46 612–625.

Schilling, M. A., P. Vidal, R. Ployhart, A. Marangoni. Learningby doing something else: Variation, relatedness, and organiza-tional learning. Management Sci. 49(1) 39–56.

Schulz, M. 2002. Organizational learning. J. A. C. Baum, ed. Com-panion to Organizations. Blackwell, Oxford, U.K.

Sine, W. D., S. Shane, D. Di Gregorio. 2003. The halo effect andtechnology licensing: The influence of institutional prestige onthe licensing of university inventions. Management Sci. 49(4)478–496.

Smith, A. 1776/1937. An Inquiry into the Nature and Causes ofThe Wealth of Nations. Edwin Cannan, ed. Random House,New York.

Song, J., P. Almeida, G. Wu. 2003. Learning by hiring: When ismobility more likely to facilitate interfirm knowledge transfer?Management Sci. 49(4) 351–365.

Sorenson, O. 2003. Interdependence and adaptability: Organiza-tional learning and the long-term effect of integration. Manage-ment Sci. 49(4) 446–463.

Starbuck, W. H. 1992. Learning by knowledge-intensive firms. J.Management Stud. 29 713–738.

Stasser, G., W. Titus. 1985. Pooling of unshared information ingroup decision making: Biased information sampling duringdiscussion. J. Personality Social Psych. 48 1467–1478.

Szulanski, G. 1996. Exploring internal stickiness: Impediments tothe transfer of best practice within the firm. Strategic Manage-ment J. 17 27–43.. 2000. The process of knowledge transfer: A diachronic anal-ysis of stickiness. Organ. Behavior Human Decision Processes 829–27.

Thomas-Hunt, M. C., T. Y. Ogden, M. A. Neale. 2003. Who’s reallysharing? Effects of social and expert status on knowledgeexchange within groups. Management Sci. 49(4) 464–477.

Thompson, L., D. Gentner, J. Lowenstein. 2000. Avoiding missedopportunities in managerial life: Analogical training morepowerful than individual case training. Organ. Behavior HumanDecision Processes 82 60–75.

Uzzi, B. 1997. Social structure and competition in interfirm net-works: The paradox of embeddedness. Admin. Sci. Quart. 4235–67.

Uzzi, B., R. Lancaster. 2003. The role of relationships in interfirmknowledge transfer and learning: The case of corporate debtmarkets. Management Sci. 49(4) 383–399.

Walsh, J. P., G. R. Ungson. 1991. Organizational memory. Acad. Man-agement Rev. 16 57–91.

Weber, M. 1922/1978. Economy and Society: An Outline of Interpre-tive Sociology. G. Roth, C. Wittich, eds. University of CaliforniaPress, Berkeley, CA.

Weber, R. A., C. F. Camerer. 2003. Cultural conflict and merger fail-ure: An experimental approach. Management Sci. 49(4) 400–415.

Wegner, D. M. 1986. Transactive memory: A contemporary analysisof the group mind. B. Mullen, G. R. Goethals, eds. Theories ofGroup Behavior. Springer-Verlag, New York, 185–205.

Williams, K. Y., C. A. O’Reilly. 1998. Demography and diversity inorganizations: A review of forty years of research. Res. Organ.Behavior 20 77–140.

Winter, S. G., G. Szulanski. 2001. Replication as strategy. Organ. Sci.12 730–743.

Zander, U., B. Kogut. 1995. Knowledge and the speed of the transferand imitation of organizational capabilities: An empirical test.Organ. Sci. 6 76–92.

Zellmer-Bruhn, M. E. 2003. Interruptive events and team knowl-edge acquisition. Management Sci. 49(4) 514–528.


argote_ mcevily_reagans_2003_managing knowledge in organizations an integrative framework and review...

Documents