Organizational Knowledge Creation Theory: A First ...kaiser/birgit/Nonaka-Papers/Organizational... · Organizational Knowledge Creation Theory Ikujiro Nonaka, * Philippe Byosiere,-i_

International Business Review Vol. 3, No. 4, pp. 337-351, 1994 Elsevier Science Ltd

Pergamon 0969-5931(94)00021-2 Printed in Great Britain 0969.5931/94 $7.00 + 0.00 1

Organizational Knowledge Creation Theory: A First

Comprehensive Test

Organizational Knowledge

Creation Theory

Ikujiro Nonaka, * Philippe Byosiere,-i_ Chester C. Borucki”f and Noboru Konnot

*Institute for Business Research, Hitotsubashi University, 2-l Naka, Kunitachi, Tokyo 186, Japan

TNijenrode University, The Netherlands Business School, Straatweg 2.5, 3621 BG Breukelen, The Netherlands

$Engine, Inc./Column, Inc., 2-12-21-301 Asagayakita, Suginami-ku, Tokyo 166, Japan

Abstract - Confirmatory factor analyses were conducted to test Nonaka’s ((1994) Organization Science, Vol. 5, No. 1, pp. 14-37) a priori model of organizational knowledge creation with data collected from 105 Japanese middle managers. The results provide strong support for viewing organizational knowledge creation as a higher-order construct comprised of four knowledge conversion processes: socialization, externalization, combination, and internalization.

Key Words - Confirmatory Factor Analyses, Organizational Knowledge Creation, Socialization, Externalization, Combination, Internalization.

Introduction Organizational capability for knowledge creation is gaining attention as a potential source of competitive advantage for firms operating in today’s global marketplace (cf. Toffler, 1990; Badaracco, 1991; Quinn, 1992). A number of firms including Nissan, NEC, Honda, Hewlett-Packard, Xerox, Motorola, 3M, and General Electric continue to hone their knowledge creation skills in a variety of areas such as developing and introducing new products, shortening design manufacturing cycle times, collaborating as well as competing with former foes, and overcoming barriers to entering new markets.

Unquestionably, the ever increasing importance of the topic calls for a shift in our thinking about the nature and functioning of organizations as knowledge creating systems. Contemporary organizational theory has long been dominated by a paradigm that conceptualizes the organization as a system that “processes” information and/or “solves” problems. Central to this paradigm is the assumption that a fundamental task for the organization is how efficiently it can deal with information and decisions in an uncertain, changing environment. This paradigm suggests that the solution lies in the “input-process-output” sequence of hierarchical information processing. Yet a critical problem with this paradigm follows from its passive and static view of the organization. Information processing is viewed as a problem-solving

337

338

International Business Review

394

activity which centers on what is given to the organization - without due consideration of what is created by it.

Any organization that dynamically deals with a changing environment should not only process information efficiently, but also create information and knowledge. Analyzing the organization in terms of its design and capability to process information imposed by the environment constitutes an important approach to interpreting certain aspects of organizational activities. However, it can be argued that the organization’s interaction with its environment, together with the means by which it creates and distributes information and knowledge, are more important when it comes to building an active and dynamic understanding of the organization (Nonaka, 1994).

A review of the literature reveals that there is a dearth of research on knowledge creation, the variables that comprise this construct, and how the process of knowledge creation can be managed to help organizations achieve and sustain competitive advantage. The study on which this paper is based aims to address some of these shortcomings. More specifically, the purpose of this paper is to build from Nonaka’s (1994) theory and propose and test an a priori higher-order model of knowledge creation. In the sections which follow, we first briefly review Nonaka’s model of knowledge creation in which relationships between epistemological and ontological dimensions of knowledge creation as well as between tacit and explicit knowledge are delineated. We then hypothesize first- and higher-order models comprised of knowledge creation factors. Subsequently, we use confirmatory factor analysis to test these models of knowledge creation with data collected from a sample of Japanese middle managers. Following a review of our results and their implications, we discuss directions for future research.

Dimensions of Knowledge Creation Although a great deal has been written about the importance of knowledge in management, relatively little is known about how knowledge is created and how the knowledge creation process can be managed. One dimension of this knowledge creation process can be drawn from a distinction between two types of knowledge identified by Polanyi (1966): tacit and explicit. Knowledge that can be expressed in words and numbers only represents the tip of the iceberg of the entire body of possible knowledge. According to Polanyi (1966), “explicit” or codified knowledge refers to knowledge that is transmittable in formal, systematic language, whereas “tacit” knowledge has a personalized quality which makes it hard to formalize and communicate. The latter deeply rooted in action, commitment, and involvement in a specific context, or as Polanyi has stated, it “indwells” in a comprehensive cognizance of the human mind and body.

While Polanyi articulates the contents of tacit knowledge in a philosophical context, it is also possible to expand his idea in a more practical direction. Tacit knowledge involves both cognitive and technical elements. The cognitive elements center on what Johnson-Laird (1983) called “mental models” in which human beings form working models of the world by

339

creating and manipulating analogies in their minds. These working models Organizational include schemata, paradigms, beliefs and viewpoints that provide Knowledge “perspectives” that help individuals to perceive and define their world. By contrast, the technical element of tacit knowledge covers concrete know-how,

Creation Theory/

crafts, and skills that apply to specific contexts. It is important to note here that the cognitive element of tacit knowledge refers to an individual’s images of reality and visions for the future, that is to say, what is and what ought to be. The articulation of tacit perspectives - in a kind of “mobilization” process - is a key factor in the creation of new knowledge.

Tacit knowledge creation is a continuous activity and embodies what Bateson (1973) has referred to as an “analogue” quality. In this context, communication between individuals may be seen as an analogue process that aims to share tacit knowledge to build mutual understanding. This understanding involves a kind of “parallel processing” of the complexities of current issues, as the different dimensions of a problem are processed simultaneously. By contrast, explicit knowledge is discrete or “digital”. It is captured in records of the past such as libraries, archives and databases and is assessed on a sequential basis.

Four Modes of Knowledge Conversion Based in part on his qualitative case studies of the knowledge creation process in a sample of Japanese companies, Nonaka (1994) identified four different patterns of interaction between tacit and explicit knowledge. These patterns represents ways in which existing knowledge can be “converted” into new knowledge. The idea of “knowledge conversion” evolves from Anderson’s ACT model (Anderson, 1983) developed in cognitive psychology. In the ACT model, knowledge is divided into “declarative knowledge” (actual knowledge) that is expressed in the form of proposition and “procedural knowledge” (methodological knowledge) which is used in such activities as remembering how to ride the bicycle or to play the piano. In the context of the present discussion, the former might approximate to explicit knowledge and the latter to tacit knowledge. Anderson’s model hypothesizes that declarative knowledge has to be transformed into procedural knowledge in order for cognitive skills to develop. This hypothesis is consistent with Ryle’s (1949) classification of knowledge into categories of knowing that something “exists” and knowing “how” it operates. Anderson’s categorization can be regarded as a more sophisticated version of Ryle’s classification. One limitation of the ACT model is the hypothesis that transformation of knowledge is unidirectional and only involves transformations from declarative to procedural knowledge, while it can be argued that transformation is bidirectional. This may be because the ACT model is more concerned with maturation than with the creation of knowledge.

The assumption that knowledge is created through conversion between tacit and explicit knowledge allows us to postulate four different “modes” of knowledge conversion from: (1) tacit knowledge to tacit knowledge, (2) explicit knowledge to explicit knowledge, (3) tacit knowledge to explicit

340

International knowledge, and (4) explicit knowledge to tacit knowledge. Below, we briefly Business review each of these conversion processes.

Review First, there is a mode of knowledge conversion that enables us to convert

374 tacit knowledge through interaction between individuals. One important point to note here is that an individual can acquire tacit knowledge without language. Apprentices work with their mentors and learn craftsmanship not through language but by observation, imitation, and practice. In a business setting, on-the-job training (OJT) uses the same principle. The key to acquiring tacit knowledge is experience. Without some form of shared experience, it is extremely difficult for people to share each other’s thinking processes. The mere transfer of information will often make little sense if it is abstracted from embedded emotions and nuanced contacts that are associated with shared experiences. This process of creating tacit knowledge through shared experience will be called “socialization”.

The second mode of knowledge conversion involves the use of social processes to combine different bodies of explicit knowledge held by individuals. Individuals exchange and combine knowledge through such exchange mechanisms as meetings and telephone conversations. The reconfiguring of existing information through the sorting, adding, and categorizing of explicit knowledge can lead to new knowledge. Modern computer systems provide a graphic example. This process of creating explicit knowedge from explicit knowledge is referred to as “combination”.

The third and fourth modes of knowledge conversion relate to patterns of conversion involving both tacit and explicit knowledge. These conversion modes capture the idea that tacit and explicit knowledge are complementary and can expand over time through a process of mutual interaction. This interaction involves two different operations. One is the conversion of tacit knowledge into explicit knowledge, which we call “externalization”. The other is the conversion of explicit knowledge into tacit knowledge, which we call “internalization” and which is similar to the traditional notion of “learning”.

Three of the four types of knowledge conversion - socialization, combination, and internalization - have partial analogs with aspects of organizational theory. For example, socialization is connected with theories of organizational culture, while combination is rooted in information processing and internalization has associations with organizational learning. By contrast, the concept of externalization is not well-developed. The limited analysis that does exist is from the point of view of information creation (cf. Nonaka, 1987).

Theories of organizational learning do not address the critical notion of externalization, and have paid little attention to the importance of socialization even though there has been an accumulation of research on “modeling” behavior in learning psychology. Another difficulty relates to the concepts of “double-loop learning” (Argyris and Schon, 1987) or “unlearning” (Hedberg, 1981), which arises from a strong orientation toward organizational development. Since the first integrated theory of organizational

341

learning presented by Argyris and Schon, it has been widely assumed, Organizational implicitly or explicitly, that double-loop learning, i.e. the questioning and Knowledge reconstruction of existing perspectives, interpretation frameworks, or decision premises, can be very difficult for organizations to implement by themselves.

Creation Theory

In order to overcome this difficulty, they argue that some kind of artificial intervention such as the use of organizational development programs is required. Seen from the vantage point of organizational knowledge creation, on the contrary, double-loop learning is not a special, difficult task but a daily activity for the organization. Organizations continuously create new knowledge by reconstructing existing perspectives, frameworks, or premises on a day-to-day basis. In other words, double-loop learning ability is “built into” the knowledge creating model, thereby circumventing the need to make unrealistic assumptions about the existence of a “right” answer.

Model Shift and Spiral of Knowledge While each of the four modes of knowledge conversion can create new knowledge independently, the central theme of the model of organizational knowledge creation proposed here hinges on a dynamic between the different modes of knowledge conversion. That is to say, knowledge creation centers on the building of both tacit and explicit knowledge and, more importantly, on the interchange between these two aspects of knowledge through internalization and externalization.

A failure to build a dialogue between tacit and explicit knowledge can cause problems. For example, both pure combination and socialization have drawbacks. A lack of commitment and neglect of the personal meaning of knowledge might mean that pure combination becomes a superficial interpretation of existing knowledge, which has little to do with here-and-now reality. It may also fail to crystallize or embody knowledge in a form that is concrete enough to facilitate further knowledge creation in a wider social context. The “sharability” of knowledge created by pure socialization may be limited and, as a result, difficult to apply in fields beyond the specific context in which it was created.

Organizational knowledge creation, as distinct from individual knowledge creation, takes place when all four modes of knowledge creation are “organizationally” managed to form a continual cycle. This cycle is shaped by a series of shifts between different modes of knowledge conversion. There are various “triggers” that induce these shifts between different modes of knowledge conversion. First, the socialization mode usually starts with the building of a “team” or “field” of interaction. This field facilitates the sharing of members’ experiences and perspectives. Second, the externalization mode is triggered by successive rounds of meaningful “dialogue”. In this dialogue, the sophisticated use of “metaphors” can be used to enable team members to articulate their own perspectives, and thereby reveal hidden tacit knowledge that is otherwise hard to communicate. Concepts formed by teams can be combined with existing data and external knowledge in a search of more concrete and sharable specifications. This combination mode is facilitated by

342

International Business Review

394

such triggers as “coordination” between team members and other sections of the organization and the “documentation” of exisiting knowledge. Through an interactive process of trial-and-error, concepts are articulated and developed until they emerge in a concrete form. This “experimentation” can trigger internalization through a process of “learning by doing”. Participants in a “field” of action share explicit knowledge that is gradually translated, through interaction and a process of trial-and-error, into different aspects of tacit knowledge.

While tacit knowledge held by individuals may lie at the heart of the knowledge creating process, realizing the practical benefits of that knowledge centers on its externalization and amplification through dynamic interactions between all four models of knowledge conversion. Tacit knowledge is thus mobilized through a dynamic “entangling” of the different modes of knowledge conversion in a process which will be referred to as a “spiral” model of knowledge creation, illustrated in Fig. 1. The interaction between tacit knowledge and explicit knowledge will tend to become larger in scale and faster in speed as more organizational actors become involved. Thus, organizational knowledge creation can be viewed as an upward spiral process, starting at the individual level moving up to the collective (group) level, and then to the organizational level, sometimes reaching out to the inter- organizational level.

Method

Subjects

The subjects in this study were 105 managers from the headquarters of a

Figure 1. Spiral of Organizational Knowledge Creation

Epistemological dimension

4 Explicit

knowledge

Tacit knowledge

Socializauon

Inrernalizalion . ‘_r Ontological

Individual dimension

Group Organization Inter-organiratron

4 Knowledge level

343

similar number of Japanese firms located in Tokyo. All of the subjects were Organizational male, occupied a middle managerial position (bucho, kacho, or kakaricho) Knowledge within their firms, and their average age was 40 years. The data were collected when the subjects were participating in a seminar at the Japan

Creation Theory

Productivity Center in May 1993.

Survey Data The data used in this study were collected through a self-administered paper and pencil questionnaire. The questionnaire was comprised of 184 items which were organized into seven major sections. Thirty-eight items were designed to operationalize the knowledge creation variables, a general description of the hypothesized factors is provided in Table 1. The survey items were measured with five-point Likert-type scales, with scale anchors ranging from a very little extent to a very great extent. The survey items were intended to measure the amount of time subjects spent on specific knowledge creation activities (called “intellectual operations”). The survey items representing each of the hypothesized knowledge creation factors were pilot tested with smaller samples in 1992. Due to space constraints, the full set of the original items is not presented.

Specification of Organizational Knowledge Creation Models A three-step procedure, following Marsh and Hocevar’s (1985, 1988) Hierarchical Confirmatory Factor Analysis method, was used to assess the validity of Nonaka’s (1994) organizational knowledge creation construct. First, confirmatory factor analyses were conducted at the item level for each of the four knowledge conversion processes. The varimax solution for each of these factor analyses is presented. Based on these results, scales were then constructed and reliability coefficients were computed for each of the scales.

The second step involved the computation of the second-order confirmatory factor analyses using the set of factors that emerged from the first-order factor analyses. The Principal Component solution yielded one factor for each of the knowledge conversion processes.

In the third step, Bentler’s (1992) EQS structural equations modeling program was used to test the third-order confirmatory factor model of knowledge creation using the four factors obtained from the second-order factor analyses. Given sample size and limited degrees of freedom, the use of EQS is appropriate for this analytic procedure as there are considerably fewer parameters estimated in the third-order four-factor model. In addition, Bentler’s EQS structural equations program enabled us to compute fit indices for Nonaka’s a priori model of knowledge creation.

Results

First-order Factor Analyses As noted above, four conversion modes of knowledge creation were identified: socialization (tacit-to-tacit knowledge), externalization (tacit-to-

344

International Business Review 3,4

Hypothesized First- and Higher-order Knowledge Creation Factors

345

I Factor 2 Factor 3 Factor 4

Socialization 5 0.83940 0.17786 -0.00746 0.057 15 6 0.73404 0.03505 0.03299 0.24125 1 0.62404 -0.0322 1 0.46002 -0.06069 8 -0.01370 0.84365 0.07378 0.18702 4 0.17601 0.78794 0.23816 0.07760 7 0.58594 0.59154 -0. I2656 0.02268 2 0.10215 0.01536 0.89937 0.05647 3 -0.02898 0.23497 0.82275 0.05879 9 0.16718 0.02349 0.03853 0.89224 10 0.04368 0.23840 0.05753 0.85291

Eigenvalue 3.12 I .57 I .41 1.189 Var explained (%) 31.2 15.7 14.1 11.9 Cum percentage (%) 31.2 46.9 61.1 73.0

Externalization 1 5 2 3 4 9 7 8 6

Eigenvalue Var explained (lo) Cum percentage (%)

Combination 1 2 3 5 8 9 IO 4 7 6 Eigenvalue Var explained (%) Cum percentage (%)

0.79579 0.77206 0.76362 0.75796 0.75185 0.72730 0.72496 0.63233 0.44306

4.60 51.1 51.1

0.85592 0.83346 0.74228 0.52417 0.00572 0.31355 0.37472 0.0263 1 0.16540

-0.14920 3.63 36.3 36.3

0.16469 -0. I2570 0.18709 0.01363

-0.02 122 0.26320 0.32174 0.3907 1 0.86274 -0.05866 0.72422 0.17719 0.68205 0.23297

-0.00635 0.78681 0.11859 0.62144 0.5476 1 0.58691

1.65 1.15 16.5 11.5 52.8 64.3

Continue1

Organizational Knowledge

Creation Theory

Table 2. First-order Factor

Analyses for Socialization,

Externalization, Combination, and

Internalization Variables Varimax

Rotated Factor Solution

346

International Business

Review 334

Table 2. Continued

Item

Internalization 9 8 4 5 7 1 2 3 6

Eigenvalue Var explained (%) Cum percentage (%)

Factor 1 Factor 2

0.79723 0.03730 0.77865 0.10564 0.63292 0.9966 0.61256 0.34513 0.59274 0.46576 0.09624 0.84438 0.18555 0.76750 0.07342 0.71938 0.40062 0.48419

3.63 1.35 40.4 15.1 40.4 55.5

Factor 3 Factor 4

explicit knowledge), combination (explicit-to-explicit knowledge) and internalization (explicit-to-tacit knowledge). The varimax solution for each of these factor analyses is presented in Table 2.

The first-order confirmatory factor analysis of ten items reflecting the socialization process (Table 2) yielded four factors explaining 73.0% of the variance. The four factors constitute the process of converting tacit knowledge to tacit knowledge. The first factor, which explains 31.2% of the variance, consists of items that pertain to the accumulation of tacit knowledge. The second factor which is comprised of extra-firm social information gathering activities, explains 15.7% of the variance. The third factor is comprised of intra-firm social information gathering activities and explains 14.1% of the variance. The final factor, which consists of items that pertain to the transfer of knowledge from the master to the different team members, explains 11.9% of the variance.

The first-order factor analysis of the nine items hypothesized to represent the externalization process (Table 2) yielded one factor explaining 5 1.1% of the variance. This finding differs from Nonaka’s theory which hypothesized that two factors - metaphor and dialogue - would be retained.

The first-order factor analysis of the ten items believed to represent the combination knowledge conversion process (Table 2) yielded three factors which explained 64.3% of the total variance. These three factors represent a

Table 3. Means. Standard Deviations, Pearson Correlation Coefficients and Cronbach Alpha Coefficients (on the

Mean Std Dev Socialization Externalization Combination Internalization

Socialization 2.744 0.674 1.00 Externalization 2.705 0.854 0.535 1.00 Combination 2.165 0.726 0.467 0.662 1.00 Internalization 2.567 0.714 0.464 0.521 0.516 1.00

Diagonal) fo; First- order Factors

Number of eases = 105

Mea

n

Std

Dev

C

OM

FA

I

CO

MF

A2

CO

MF

A3

EX

TF

A13

IN

TF

A 1

INT

FA

2 S

OC

FA

l S

OC

FA

2 S

OC

FA

3 S

OC

FA

4

CO

MF

A

I 2.

93 1

1.

074

(0.7

86)

CO

MF

A2

3.03

9 0.

964

0.36

6

CO

MF

A3

2.31

3 0.

832

0.29

5

EX

TF

Al

2.70

6 0.

839

0.50

4

INT

FA

1

2.66

7 0.

812

0.21

2

INT

FA

2 2.

461

0.86

4 0.

323

SO

CF

A

1 2.

461

0.97

2 0.

154

SO

CF

A2

2.63

7 0.

983

0.28

6

SO

CF

A3

3.00

9 0.

917

0.38

3

SO

CF

A4

3.00

9 0.

999

0.15

7

Nu

mbe

r of

cas

es =

IO

2

(0.7

64)

0.41

4

0.54

1

0.33

3

0.45

4

0.24

5

0.26

6

0.27

9

0.34

9

(0.6

02)

0.49

9 (0

.875

)

0.31

8 0.

392

(0.7

13)

0.37

9 0.

544

0.46

1 (0

.766

)

0.21

4 0.

277

0.37

2 0.

216

(0.7

00)

0.15

5 0.

409

0.13

2 0.

245

0.37

4 (0

.785

)

0.25

5 0.

371

0.24

4 0.

357

0.11

7 0.

202

(0.6

79)

0.16

2 0.

275

0.51

6 0.

247

0.40

3 0.

185

0.30

2 (0

.752

)

348

International Business Review

3,4

three step sequence of data processing. The first factor, which explains 36.3% of the variance, is comprised of items pertaining to the acquisition and integration of information. The third factor which represents the second step in the combination process (the synthesis and processing of the information collected in the first step), explains 11.5% of the variance. The second factor, which is actually the final step in the hypothesized data processing sequence, consists of items that pertain to the dissemination of the information collected in step one and put into structure in step two. This factor explains 16.5% of the variance.

The first-order confirmatory factor analysis of the nine items believed to represent the internalization process (Table 2) yielded two factors, as hypothesized, which explain 55.5% of the total variance. The first factor consists of items that primarily reflect internalization through personal experience and explains 40.4% of the variance. The second factor, which explains 15.1% of the variance, is comprised of items that pertain to internalization through experimentation and simulation, which is similar to the acquisition of “virtual” knowledge. Table 3 presents the means, standard deviations, Pearson correlations and Cronbach alpha coefficients (on the diagonal) for the factors obtained from the first-order factor analyses. These factors are used in the second-order factor analyses.

Second-order Factor Analyses As hypothesized, the second-order factor analysis conducted for each of the knowledge conversion processes yielded one overall factor in each instance. Owing to space constraints, results of these factor analyses are not presented in this paper.

Means, standard deviations, Pearson correlations for the factors obtained in the second-order factor analysis are presented in Table 4. These factors serve as the higher-order factors in the structural equation analysis discussed in the next section.

Higher-order Factor Model The third-order four-factor model, as depicted in Fig. 2, was estimated with Bentler’s EQS structural equation program. The model yields high goodness- of-fit indices (NFI = 0.989; NNFI = 1 .OO; CFI = 1 .OO), and the chi square value is highly non-significant (P = 0.447, which indicates that the model fits the data as hypothesized very well.

Discussion Recent developments in organizational theory suggest that organizational performance in terms of innovation, product development and competitive advantage is highly determined by the creation of knowledge at the individual, group and organizational levels (Nonaka, 1994). Building from concepts introduced by Polanyi (1966), Nonaka’s (1994) dynamic theory of knowledge creation identifies four major knowledge conversion modes which are based on the interchange between tacit and explicit knowledge. In this

349

ihoael malces Chi square = 1.608 df = 2 P = 0.447 Bentler-Bonnett NFI = 0.989 Bentler-Bonnett NNFI = 1 .OO Comparative fit index = 1 .OO

study, based on a sample of 105 Japanese middle managers, we tested and validated the hypothesis that the construct of knowledge creation consists of four major knowledge conversion processes: socialization, externalization, combination, and internalization. All four knowledge conversion factors or processes explain a high amount of variance in the organizational knowledge creation construct. The first- and second-order factor analyses suggest that each of the four knowledge conversion processes consists of specific subfactors. More specifically, socialization, whereby tacit knowledge is converted to tacit knowledge, is comprised of four subconstructs: the mere accumulation of tacit knowledge, the gathering of social information internal to the organization (also called “wandering inside”), the gathering of social information outside the organization in a broader societal context (“wandering outside”), and finally the transfer of tacit knowledge from the master to the team members.

The knowledge conversion process where tacit knowledge is converted to explicit knowledge, also known as externalization, consisted of one large factor in this study, though two factors were hypothesized. Combination, which is the process by which explicit knowledge is converted into explicit knowledge, consists of three steps which conform to what happens in most organizations today: the integration, the synthesis and dissemination of existing knowledge into better structurally organized forms for broad dissemination. Finally, explicit knowledge in the organization may be

Organizational Knowledge

Creation Theory

Figure 2. Higher-order

Structural Equation Model: Estimates

and Statistics

350

International Business Review

374

converted into tacit knowledge in two different ways: through personal experience in which knowledge is acquired from real world experiences, and through simulation and experimentation in which knowledge is acquired from the virtual world.

Conclusion Using a combination of classical multivariate confirmatory factor analysis and structural equation modeling techniques, we were able to test the four underlying modes of knowledge conversion as suggested in Nonaka’s dynamic theory of organizational knowledge creation. The main conclusion of this study is that as suggested by Nonaka (1994), organizational knowledge creation consists of four underlying constructs representing each of the four knowledge conversion modes. In contemporary organizations, enhanced organizational performance is frequently attributed to the conversion of explicit knowledge to explicit knowledge, or what is referred to as combination in Nonaka’s theory. However, as the results of the present study suggest, the three other knowledge conversion processes, all of which involve tacit knowledge, play an important role in the organizational knowledge creation process.

This study has certain limitations. First, this was the first time the knowledge creation questionnaire was administered for non-pilot testing purposes. The convenience sample of 105 Japanese middle managers was rather small, which constrained us from being able to test a full-scale structural equation model starting at the item level. Second, the sample is heterogeneous in that each respondent represented a different Japanese firm, which raises questions regarding internal validity. Relatedly, given that all the data were collected in Japan, generalizability of these findings to other cultures remains questionable, at best. Third, the study would have been strengthened if qualitative data were available to enrich the empirical findings.

Suggestions for future research include the following. First, the theory testing in this study focused primarily on the content of organizational knowledge creation. In a next phase it is important to investigate the process of organizational knowledge creation. More specifically, future research should examine the extent to which the four modes of knowledge creation are related and if the causal sequence hypothesized in the form of Nonaka’s (1994) spiral model is manifested. Also warranting closer attention are levels of analysis issues and how knowledge creation is enhanced or hindered as one moves from the individual to the inter-organizational or industry levels. Relatedly, future research also calls for testing the theory within a single, large organizational setting with a relatively homogenous population. Cross- national comparisons ought to be undertaken to also investigate the extent to which the four knowledge conversion modes occur in North American and European organizations. Furthermore, industry specific comparison ought to be looked at. It may well be the case that certain industries rely more on tacit knowledge conversion or explicit knowledge conversion than others. Another

351

issue that remains is to examine the extent to which external forces such as Organizational economic conditions and global competition as well as internal factors such as Knowledge strategy, organizational culture, human resources practices, organizational structure, and technology impact the organizational knowledge creation

Creation Theory

process. In summary, the results of this study confirm a four-factor higher-order

model of organizational knowledge creation, The primary contribution of this study is the operationalization and the empirical testing of the theory of organizational knowledge creation. We feel very strongly that in the next decade organizations will be required to rely more on tacit knowledge present in the organization in order to achieve and sustain global competitive advantage. The theory of knowledge creation adds another little-investigated dimension to contemporary organization theories such as open-systems theory and the population ecology paradigm, among others.

References Anderson, J.R. (1983) The Architecture of Cognition. Harvard University Press, Cambridge,

Massachusetts. Argyris, C. and Schon, D.A. (1987) Organizational Learning. Addison-Wesley, Reading,

Massachusetts. Badaracco, J. (1991) The Knowledge Link; Competitive Advantage through Strategic Alliances.

Harvard Business School Press, Boston, Massachusetts. Bateson, G. (1973) Steps to an Ecology ofMind. Paladin, London. Bentler, P.M. (1992) E&S Structural Equations Program A4unual. BMDP Statistical Software,

Los Angeles, California. Hedberg, B.L.T. (1981) How Organizations Learn and Unlearn, in Nystrom, PC. and Starbuck

W.H. (Eds), Handbook of Organizational Design. Oxford University Press, Oxford. Johnson-Laird (1983) Mental Models. Cambridge University Press, Cambridge. Marsh, H.W. and Hocevar, D. (1985) The Application of Confirmatory Factor Analysis to the

Study of Self-concept: First- and Higher-order Structures and their Invariance Across Age Groups. Psychological Bulletin, Vol. 97, pp. 562-582.

Marsh, H.W. and Hocevar, D. (1988) A New, More Powerful Approach to Multitrait- multimethod Analyses: Application of Second-order Confirmatory Factor Analysis. Journal of Applied Psychology, Vol. 73, No. 1, pp. 107-I 17.

Nonaka, I. (1987) Managing the Firm as Information Creation Process, in Meindl, J. et ul. (Eds), Advances in Information Processing in Organizations, Vol. 4, 1991, pp. 239-275. JAI Press, Greenwich, Connecticut.

Nonaka, I. (1994) A Dynamic Theory of Organizational Knowledge Creation. Organization Science, Vol. 5, No. 1, February, pp. 14-37.

Polanyi, M. (1966) The Tacit Dimension. Routledge & Kegan Paul, London. Quinn, J.B. (1992) Zntelligent Enterprise. Free Press, New York. Ryle, G. (1949) The Concept ofMind. Huchison, London. Tofler, A. (1990) Powershift: Knowledge, Wealth and Violence at the Edge of the 2 I st Centmy.

Bantan Books, New York.