One of the most serious threats to the relatively new field of
strategic management research is poor construct measurement (Boyd et
al., 2005). Due to the relatively complex nature of strategic management
variables, quality of measurement is crucial (Godfrey and Hill, 1995)
and strategic management research must place greater emphasis on
research design, construct validation, and more sophisticated analytical
techniques (Bergh, 2001). Many researchers tend to operationalize latent
constructs with the use of proxy variables. For example, Boyd et al.
(2005) found that constructs such as available organizational resources,
public profile, core rigidity, and ability to initiate competitive
action have all been operationalized by single-item archival measures of
organizational size. Additionally, organizational size can be measured
as number of employees, number of products, number of production
facilities, or any of at least a dozen other measures. The reliance on
single-item measures to the exclusion of multi-item scales virtually
ensures that research is conducted with unreliable measures that
attenuate results. More importantly, there is a clear question of
validity. Does organizational size truly represent core rigidity,
ability to initiate competitive action, or something altogether
different? Additionally, Payne et al. (2003) suggest that the use of
archival data has several disadvantages. These include the fact that the
data may have been originally collected for some other purpose, there
may be missing data points, archival data may be susceptible to
experimenter bias as some researchers might examine the data before they
propose hypotheses, and that archival data may not be readily
analyzable. Given that all of the above are distinct problems plaguing
strategic management research, the truly important issue is that there
is no real valid representation of many key constructs.
In fact, Boyd et al. (2005) call for the development of new
multi-indicator measures for constructs in strategic management
research, especially for more contemporary approaches to strategy.
Traditional research in strategic management has used a variety of
measures for the Miles and Snow (1978) typology of prospectors,
analyzers, defenders, and reactors, and the Porter (1980) typology of
low cost, differentiation, and focus. However, these approaches to
strategic management have been criticized for a variety of reasons (see
Grant, 2005). Today, the school of strategy that is attracting a wide
degree of attention is the knowledge-based view of the firm, with its
typology of exploration and exploitation knowledge strategies. However,
an acceptable multi-indicator measure of these constructs has not been
generally accepted in the literature.
Thus, the purpose of our study is to examine the factor structure
of responses to items designed to measure the knowledge strategy
constructs of exploration and exploitation and to provide evidence of
external validity using other measures of theoretically-related
variables. In keeping with this concern for greater methodological rigor
we suggest that the further delineation of these two constructs is
appropriate and necessary for future research in this area. In the
sections that follow, we give some insight into the theoretical
background of our focal constructs, an overview of our analytical
methods, the results of our analysis, and a discussion of these results.
THEORETICAL BACKGROUND
A knowledge strategy can be viewed as a firm's set of
strategic choices regarding two knowledge domains: 1) exploration, or
the creation or acquisition of new knowledge and 2) exploitation, or the
ability to leverage existing knowledge to create new organizational
products and processes. A firm's knowledge strategy guides its
resource allocation--the degree to which the firm focuses its resources
on either generating radically new knowledge or incrementally enhancing
the existing knowledge base (March, 1991; Bierly and Chakrabarti, 1996).
In addressing these trade-offs, March (1991) argued that exploitation is
likely to maximize profits in the short run, and that exploration is
more likely to maximize long-term firm success. Accordingly, exploration
entails higher costs and increased risk for a firm, but is more likely
to lead to a sustainable competitive advantage. However, concentrating
resources too heavily on exploration may prevent firms from reaping the
benefits that come from developing these knowledge breakthroughs.
Focusing on exploration tends to slow down the development and
refinement of skills and processes associated with the firm's
current competencies. On the other hand, a strong commitment to an
exploitation strategy entails trade-offs as well. According to March
(1991), organizations that focus on the incremental knowledge gain
associated with exploitation may find themselves to be experts in areas
that have become obsolete, thus getting better and better at things that
customers no longer value.
Exploration and Exploitation Constructs
Conceptualizing exploration and exploitation as two separate
constructs implies that they are not simply the two extremes of a single
continuum (wherein movement toward one strategy inherently means
movement away from the other). Rather, they are two sets of strategic
choices, each positioning the firm to develop their intellectual capital
in a specific direction, toward excellence in either the creation and
acquisition of new knowledge or the leveraging of existing knowledge. In
essence, the orthogonal nature of the constructs indicates that firms
may pursue one, both (simultaneously), or neither of the strategies.
Because certain strategic choices, such as the decision to develop
competence in a specific technology, may be necessary to both
exploration and exploitation, and therefore would represent overlap in
the two sets of strategic choices, we expect a moderate degree of
positive correlation between the two constructs. Researchers (March,
1991; Bierly and Chakrabarti, 1996; Zack, 1999) have made claims as to
the existence of differentiated knowledge strategies; however, there has
been very little empirical research conducted in this area. Using a
cluster analysis, Bierly and Chakrabarti (1996) did find evidence of
distinct knowledge strategies, but their study was limited to large
resource-rich pharmaceutical companies. Yet we believe each orientation
to have enough unique elements that exploration can be viewed as
independent and distinct from exploitation.
The general argument made by many of the researchers mentioned
above is that choosing between exploration and exploitation necessitates
trade-offs, and therefore the two strategies are substitutes and, thus,
negatively correlated. According to this view, firms that develop
competencies in exploitation are likely to focus more of their resources
on further exploitation and fewer resources on exploration and vice
versa (Levinthal and March, 1993). Basically, this view stresses the
value of specialization and efficiency of learning. An alternative
perspective, proposed by Knott (2002), is that exploration and
exploitation are complements rather than substitutes. Firms that develop
the capabilities necessary to foster exploration are also more likely to
engage in exploitation, and vice versa. Even though there are
difficulties in simultaneously pursuing exploration and exploitation (as
described above), there are also organizational systems and human
resource practices that support both. More specifically, team-based
structures, an organizational culture that values and promotes change,
open communication channels, and human resource practices that promote
creativity and innovation will help to sustain both exploration and
exploitation (Bierly and Daly, 2002). Consistent with this perspective,
Helfat and Raubitschek (2000) argue that successful exploitation
(referred to as incremental learning) in the past can lead to and
support exploration (referred to as step function learning), and that
current exploration promotes future exploitation. Anecdotal evidence
that some companies can successfully pursue exploration and exploitation
simultaneously is provided by Knott's (2002) case study of Toyota,
Ichijo's (2002) study of General Electric, and Helfat and
Raubitschek's (2000) studies of Sony, Canon, and NEC.
Consistent with Knott's (2002) perspective, we conceptualize
explorer and exploiter as independent constructs and have developed two
four-item scales to measure individual's perceptions of the
knowledge strategy constructs known as exploration and exploitation.
Even though the concepts of exploration and exploitation are central
components of the popular knowledge-based view of the firm, little
research has been conducted to examine the legitimacy and relationship
of these constructs. An acceptable measurement tool has not yet been
developed that has been properly field tested and widely accepted. The
vast majority of articles about exploration and exploitation are either
conceptual, use computer simulations (e.g., March, 1991; Lee et al.,
2003; Garcia et al., 2003), or are based on case studies (e.g. Knott,
2002; Helfat and Raubitschek, 2000; Holmqvist, 2004; McNamara and
Baden-Fuller, 1999). A few researchers use patent, or research and
development data, or they focus on the number of new products from a
firm as proxies for exploration and exploitation (e.g., Bierly and
Chakrabarti, 1996; Katila and Ahuja, 2002; Rosenkopf and Nerkar, 2001).
To our knowledge, only He and Wong (9004) have used perceptual measures
gathered in a survey to measure these specific constructs.
Radical and Incremental Innovation
Researchers (e.g., Volberda, 1996; Hedlund, 1994) have pointed out
that explorers and exploiters often require very different types of
organizational cultures, competencies, and structures. Therefore, once a
firm creates a competency in either exploration or exploitation, it is
usually more efficient for the firm to continue on that particular path
(Levinthal and March, 1993). Additionally, researchers in the field of
management of technology have discussed the difference between radical
and incremental innovations (Damanpour, 1991; Dewar and Dutton, 1986;
Ettlie et al., 1984), which can be viewed as outputs of exploration and
exploitation, respectively. Thus, a valid measure of an
exploration-based knowledge strategy orientation should be associated
with distinctive competencies that are required to develop and implement
a radical innovation. Specifically, explorers should be firms that
aggressively invest in research and development, are creative in
improving product technologies, and have established new product
development processes that enhance their ability to bring new products
to market quickly. On the other hand, a valid measure of an
exploitation-based knowledge strategy orientation should be associated
with distinctive competencies that are required to develop and implement
an incremental innovation. Exploiters should invest more in new process
technologies to reduce their cost structure, and excel at practices that
facilitate customer satisfaction and promote continuous improvement,
such as total quality management (TQM) and benchmarking. With these
conceptualizations in mind we suggest:
H1: Distinctive competencies associated with radical innovation
will be more strongly related to an explorer orientation than to an
exploiter orientation.
H2: Distinctive competencies associated with incremental innovation
will be more strongly related to an exploiter orientation than to an
explorer orientation.
METHOD
Participants
Samples were drawn from employees of small to mid-size
manufacturers located in the mid-Atlantic region of the United States.
The firms represent a broad spectrum of manufacturing, belonging to 18
different Standard Industrial Classification (SIC) groups, including:
Food Products (9 firms), Lumber and Wood Products (10), Printing and
Publishing (5), Mechanical and Computer Equipment (8), Measuring,
Analyzing and Controlling Instruments (5), Furniture and Fixtures (5),
Chemical Products (2), Rubber and Plastics (4), Electronic Equipment
(6), Primary Metal Industries (5), Transportation Equipment (3),
Fabricated Metal Products (9), Stone and Concrete Products (2), Paper
Products (2), Textile Mill Products (1), Apparel (3), Petroleum Refining
and Related Products (1), and Miscellaneous (18).
Participating firms were identified through cooperation with state
Small Business Development Centers which, using membership lists and
small business directories, provided company names and contacts. Almost
90% of these companies met the Small Business Association criteria for
definition of a small business, having 500 or fewer employees, with the
remaining companies classified as mid-size firms (having between 501 and
1,800 employees). Of firms that were identified as subsidiaries of
larger parent companies, only those that operated as independent profit
centers (strategic business units) within the larger organization were
included. Three surveys were sent to 250 companies that initially agreed
to participate in the study. Contacts were asked to give surveys to
three individuals working in different positions within the company. To
ensure confidentiality and encourage participants to be candid in their
responses, each respondent enclosed and sealed their survey in an
envelope before returning it to the contact person at their company.
As Sharfman (1998) suggests in his study regarding the use of CEOs
as sole informants in strategy research, prudent researchers should not
rely solely on CEOs, but rather that more informants provide a richer
picture of strategy constructs than one informant. Thus, we collected
data from a variety of firms using respondents in a variety of positions
within those firms. Ninety-eight different companies returned complete
and usable surveys from three different respondents in their firms for a
response rate of 39.2%. Sample One was comprised of respondents with the
following positions in the 98 firms: 9.2% were CEO or President, 18.4%
were Human Resources representatives, 33.7% were in Productions or
Operations, 6.1% were in Sales or Marketing, 26.5% were in Finance,
Accounting, or Administration, and 6.1% were something other. Sample Two
was comprised of respondents with the following positions in the firms:
9.2% were CEO or President, 15.3% were Human Resources representatives,
42.9% were in Productions or Operations, 8.2% were in Sales or
Marketing, 16.3% were in Finance, Accounting, or Administration, and
8.2% were something other. Sample Three was comprised of respondents
with the following positions in the firms: 11.2% were CEO or President,
13.3% were Human Resources representatives, 38.8% were in Productions or
Operations, 10.2% were in Sales or Marketing, 19.4% were in Finance,
Accounting, or Administration, and 7.1% were something other. Each of
these samples is remarkably similar with respect to respondents'
positions in the 98 firms and they provide a reasonable cross-section of
employees.
Measures
The surveys completed by respondents contained a number of items
designed to elicit information about their individual perception of
their firm's knowledge strategy orientation and distinctive
competencies. Distinctive competencies were selected as potential
correlates so that an effort at external validation could be undertaken.
Operationalizations of the variables are provided below.
Knowledge Strategy Orientation. Participants completed information
on their perception of their employing firm's knowledge strategy
orientation via two sub-scales assessing the Explorer Orientation and
the Exploiter Orientation. A five-point Likert type scale anchored by 1
= strongly disagree and 5 = strongly agree formed response scales for
both sets of items. The explorer scale items focus on the key elements
frequently associated with this construct in the literature: newness,
radicalness, and creativity of ideas, technologies and products (March,
1991; Levinthal and March, 1993; Bierly and Chakrabarti, 1996; Zack,
1999). The items are: (1) We frequently experiment with radical new
ideas (or ways of doing things), (2) At our company, employees
frequently come up with creative ideas that challenge conventional
ideas, (3) Compared to our principal competitors, a high percentage of
our company sales come from new products launched within the past three
years, and (4) We are usually one of the first companies in our industry
to use new, breakthrough technologies. Cronbach's alpha for scores
on this sub-scale was .71 in Sample One, .74 in Sample Two, and .70 in
Sample Three.
The exploiter scale items capture the key elements frequently
associated with this construct in the literature: refinement and
extension of current technologies and products, increasing efficiency,
and improving procedures (Holmqvist, 2004; March, 1991; Levinthal and
March, 1993; Bierly and Chakrabarti, 1996; Zack, 1999). The items are:
(1) Our company excels at continually improving our existing products,
(2) At our company, a strong emphasis is placed on improving efficiency,
(3) Our company excels at refining existing technologies, and (4) We
frequently adjust our procedures, rules and policies to make things work
better. Cronbach's alpha for scores on this sub-scale was .72 in
Sample One, .70 in Sample Two, and .75 in Sample Three. Together, the
eight items measuring Exploration and Exploitation comprise our
Knowledge Strategy Orientation Scale (KSOS).
Distinctive Competencies. As early as 1980 Snow and Hrebiniak used
a firm's expertise in various functional areas as a measure of
distinctive competencies. In our assessment we focused on the degree of
expertise evidenced in the functional areas of the organization and
adapted several items from Delaney and Huselid (1996). Participants were
asked to assess their organization's level of expertise in the
following eight competency areas. The competencies associated with
radical innovation include: basic research and development, product
technologies, new product development, and speed in bringing a new
product to market. The competencies associated with incremental
innovation include: process technologies, knowledge of customer
preferences, cost reduction, and benchmarking. For each of these
distinctive competencies, participants were asked to indicate, on a
scale of 1 to 10, where their firm ranked as compared to their main
competitors (with 1 being well below industry average, 5 being the
industry average, and 10 being well above the industry average).
Although the relationships between our focal constructs and distinctive
competencies are analyzed at the item level of distinctive competencies,
evidence of the internal consistency of the items associated with
radical innovation is provided by Cronbach's alpha of .85 in Sample
One, .83 in Sample Two, and .85 in Sample Three. Cronbach's alpha
for the items designed to measure incremental innovation was .87 in
Sample One, .79 in Sample Two, and .85 in Sample Three. Thus, the
distinctive competencies associated with radical and incremental
innovation appear to be measuring their intended constructs.
Analysis
Exploratory Factor Analysis. In order to pre-test our items and
explore the underlying factor structure of our knowledge strategy
orientation subscales, we used principal axis factoring in an
Exploratory Factor Analysis (EFA), with a promax rotation, on the data
from Sample One. While there is some disagreement on minimum sample size
requirements, many methodologists suggest at least five to ten
respondents per item are needed for EFA (Comrey, 1988; Hair et al.,
1998). We have more than 12 respondents per item. We used the latent
root criteria of an eigen value greater than one and a scree plot for
the determination of factor extraction. Additionally, we considered
items with loadings of greater than .40 to be "substantial"
(Floyd and Widaman, 1995) and loadings above .50 to be "very
significant" (Hair et al., 1998). Because we have a theoretical
basis to support our belief that these constructs are correlated, we
used the promax form of oblique rotation. It must be noted that EFA
tends to capitalize on the chance characteristics of a sample. Because
the purpose of this article is to examine the factor structure of
responses to our scale items, we later used confirmatory factor analysis
to cross-validate the results of our Sample One EFA.
Confirmatory Factor Analysis. Confirmatory factor analysis allows a
confirmatory, rather than an exploratory, approach to determining the
underlying structure of observed variables (Harris and Schaubroeck,
1990), and provides a means of assessing the relationships between
constructs without the bias commonly introduced by measurement error
(Judd et al., 1986). Confirmatory factor analysis is used to determine
the extent to which alternative models explain the relationships between
items in a scale. Two competing measurement models of strategic
orientation were evaluated in this study. The alternative CFA models
were: (a) a one-factor model that forced all items designed to measure
Explorer and Exploiter onto a single factor of Knowledge Strategy
Orientation and (b) a two-factor model that forced the Explorer items
and the Exploiter items onto separate factors. In each model, error
variances for the items were not allowed to correlate. Should the
one-factor model provide a fit of the data equivalent to the two-factor
model, it would indicate a single underlying latent construct (i.e.,
Explorer and Exploiter as opposite ends of an unidimensional knowledge
strategy continuum). If the two-factor model should provide the better
fit than the one-factor model, then our conceptualization of Explorer
and Exploiter as distinct and independent constructs will be supported.
As suggested by Thompson and Daniel (1996), CFA is most useful when
the researcher tests a priori models, because more effective decisions
can then be made about the viability of the target model. Because the a
prior/models above are nested, the chi-square difference can be used to
test for significant differences between the models. If the chi-square
difference is significant, it indicates that the more complex two-factor
model fits the data significantly better than the simpler one-factor
model.
Hu and Bentler (1998, 1999) recommend that several goodness-of-fit
tests be conducted and that their resulting indices be reported. These
indices are of two types: absolute and incremental. An absolute index
tests how well the model covariance matrix reproduces the sample
covariance matrix while an incremental index tests the fit of the
hypothesized model as compared to a baseline model. The most
commonly-used absolute fit index is the chi-square test that assesses
the discrepancy between the implied covariance matrix of the
hypothesized model and the sample covariance matrix. A non-significant
chi-square is the desired result of this test as it suggests the model
may be a reasonable approximation of the data. However, many researchers
(c.f. Fan et al., 1999; Hu and Bentler, 1995) have cautioned that using
the chi-square test as an assessment of fit can be confounded by sample
size because as sample size increases, the chance of the chi-square test
supporting a fit of the data decreases. Thus, small differences between
the sample covariance matrix and the reproduced covariance matrix may be
determined to be statistically significant and lead to rejection of the
model. With this in mind, supplemental absolute indices were employed.
Another absolute index, the standardized root mean square residual
(SRMR) is reported as a summary statistic based upon residuals between
the elements of the implied and observed covariance matrices. The
standardized root mean square residual ranges from 0 to 1 and values
close to 0 are preferred. In fact, Hu and Bentler (1998, 1999) suggest
that researchers always use the SRMR to assess model fit because of its
sensitivity to simple model misspecification (misspecified factor
correlations). They suggest that target values of the SRMR should be
less than .08 in order to indicate adequate model fit. Another absolute
fit index, the root mean square error of approximation (RMSEA), is
reported in this study as well. The RMSEA assesses lack of fit based
upon model misspecification and provides a measure of this discrepancy
per degree of freedom (Browne and Cudeck, 1993). This fit index is quite
sensitive to complex misspecification (i.e., misspecified factor
loadings; Hu and Bentler, 1998). It ranges from 0 to 1, with target
values of less than .08 indicating adequate fit (Browne and Cudeck,
1993).
Incremental fit indices are also recommended (Hoyle and Panter,
1995; Hu and Bender, 1999) to assess model fit. The comparative fit
index (CFI) developed by Bender (1990) is reported here. It is sensitive
to misspecified factor loadings (Hu and Bentler, 1998) and assesses the
improvement of fit of the hypothesized model over the null model. The
null model is an independence model in which variables are hypothesized
to be uncorrelated. The CFI ranges from 0 to 1, and values greater than
.95 have recently been advocated (Hu and Bender, 1999) as an increase
from earlier target values greater than .90 (Hoyle and Panter, 1995).
RESULTS
Sample One EFA Results
In Sample One, our principal axis analysis resulted in two factors
with eigen values of 3.409 and 1.086 being extracted that explained
56.15% of the variance. As we envisioned, our promax oblique rotation
resulted in each Explorer item loading more highly on one factor than
the other and each Exploiter item loading more highly on the other
factor. Three of four Explorer items showed "very significant"
loadings greater than .60. Two of four Exploiter items showed
"substantial" loadings greater than .40, while another item
showed "very significant" loading. See Table 1 for the
resulting pattern matrix. With this factor structure in mind we then
cross-validated these results on the data from Samples Two and Three
using CFA.
Item Level Statistics for Samples Two and Three
Each CFA measurement model was estimated in this study using LISREL
8.71 software (Joreskog and Serbom, 2004). A component of the LISREL
software, PRELIS 2.30, was used to assess univariate normality and to
generate the covariance matrix upon which the CFA was conducted. Kline
(1998) advocates upper boundaries of 3.0 for skewness and 8.0 for
kurtosis as indicators of univariate normality.
Sample Two. The univariate data for the Explorer and Exploiter
scales were approximately normally distributed with skewness for the
eight manifest indicators ranging from -0.88 to 0.39, and kurtosis
ranging from -1.56 to 1.45 (see Table 2). Based upon the descriptive
statistics for the sample, it appears that the data were approximately
normally distributed. Therefore, the maximum likelihood (ML) method of
estimation was employed in CFA.
Sample Three. The univariate data for the Explorer and Exploiter
scales were approximately normally distributed with skewness for the
eight manifest indicators ranging from-1.10 to 0.44, and kurtosis
ranging from -1.39 to 1.31 (see Table 2). Based upon the descriptive
statistics for the sample, it appears that the data were approximately
normally distributed. Therefore, the maximum likelihood (ML) method of
estimation was employed in CFA.
Confirmatory Factor Analysis Results
The eight items comprising the Explorer Orientation and Exploiter
Orientation scales were subjected to CFA. Two models were compared: a
one-factor model forcing all eight items onto the same factor and a
two-factor model forcing the Explorer items and Exploiter items onto
their respective factors. Error terms were not allowed to correlate in
either model, and in the two-factor model items were not allowed to
cross-load. See Table 3 for the fit indices of these two models in our
samples.
Sample Two. The most complex model was the two-factor model, which
resulted in CFI = 0.92, RMSEA = 0.097, and SRMR = 0.065. The SRMR
indicates good fit of the data to the model, and the RMSEA and SRMR are
only slightly outside the recommended thresholds. The more parsimonious
one-factor model resulted in CFI = 0.79, RMSEA = 0.16, and SRMR = 0.098.
None of these indices meets the criteria for good fit. Additionally, the
[chi square] for the two-factor model was 39.60 (p < .001), while the
[chi square] for the one-factor model was 73.24 (p < .001), resulting
in a [DELTA][chi square] of 33.64 (p < .001). The significant
[DELTA][chi square] indicates that the two-factor model fits the data
significantly better than the one-factor model, providing evidence of
the superior fit of the two-factor model.
Kline (1998) suggests that evidence of convergent validity is
provided when items specified to measure a construct all have relatively
high path coefficients in CFA analysis. He further states that evidence
of discriminant validity is provided when intercorrelations between
constructs are not excessively high (i.e., > .85). In Sample Two, the
standardized path coefficients for the Explorer scale range from .40 to
.80, while the standardized path coefficients for the Exploiter scale
range from .39 to .68, thus providing some evidence of convergent
validity. Confirmatory factor analysis reveals that the disattenuated
correlation between the Explorer factor and the Exploiter factor was
.59, thus providing some evidence of discriminant validity between the
constructs. In an effort at further validating the factor structure of
the data, we submitted data from Sample Three to CFA.
Sample Three. Before we could assess the relationship between our
Explorer and Exploiter constructs in Sample Three, we assessed the fit
of the two-factor model on data from that sample. The two-factor model
resulted in CFI = 0.95, RMSEA = 0.093, and SRMR = 0.066. Of these
indices, the CFI and SRMR indicate good fit of the data to the model,
but the RMSEA is slightly outside the recommended range. The more
parsimonious one-factor model resulted in CFI = 0.91, RMSEA = 0.120, and
SRMR = 0.080. Of these, only the SRMR meets the criteria for good fit.
Additionally, the [chi square] for the two-factor model was 35.43 (p
< .001), while the [chi square] for the one-factor model was 48.28 (p
< .01), resulting in a [DELTA][chi square] of 12.85 (p < .001).
The significant [DELTA][chi square] indicates that the two-factor model
fits the data significantly better than the one-factor model, providing
yet more evidence of the superior fit of the two-factor structure.
The standardized path coefficients for the Explorer scale range
from .54 to .73, while the standardized path coefficients for the
Exploiter scale range from .51 to .75, thus providing some evidence of
convergent validity (see Table 4). Confirmatory factor analysis reveals
that the disattenuated correlation between the Explorer factor and the
Exploiter factor was .67, thus providing some evidence of discriminant
validity. See Table 4 for the completely standardized factor pattern and
squared multiple correlations for the two alternative models in both
samples.
Relationship to Distinctive Competencies
We used a Z-test described by Meng, Rosenthal, and Rubin (1992) to
examine our hypotheses that test for significant differences between
correlated correlation coefficients. Each of our distinctive
competencies is correlated with each of our focal constructs: Explorer
and Exploiter.
In Sample Three, two of the four distinctive competencies
associated with radical innovation were statistically stronger in their
relationship with Explorer than with Exploiter. The significant
correlations were between new product development (Z = 2.152, p <
.05) and speed to market (Z = 2.344, p < .05) in their relationships
with Explorer. Z-scores were 0.454 and -0.499 for research and
development and product technology, respectively. Thus, there was
partial support for hypothesis one in sample three. Additionally, each
of the distinctive competencies associated with incremental innovation
were statistically stronger in their relationship with Exploiter than
with Explorer. The correlations were between process technology (Z =
-2.021, p < .05), customer preferences (Z = -2.468, p < .05),
operating efficiency (Z = -3.048, p < .05), and benchmarking (Z =
-2.148, p < .05) in their relationships with Exploiter. Thus, there
was support for hypothesis two in Sample Three. See Table 5 for the
correlations and accompanying Z tests.
DISCUSSION
This study examines the individual perceptions of employees
regarding the knowledge strategy orientation of their firms. More
specifically, we are interested in whether these individuals perceive
their organization to have an Explorer or an Exploiter knowledge
strategy orientation.
Measurement Implications
We analyzed the self-report data in three samples of respondents
using EFA and CFA to examine the factor structure of item responses, the
convergent and discriminant validity of the explorer and exploiter
constructs, and the relationship of these measures with the measures of
distinctive competencies. Our results indicate that our scales are
indeed separate factors, as a one-factor model did not fit the data
well, but the two-factor model did. The Explorer and Exploiter scales
also resulted in scores with acceptable reliability. The factor pattern
coefficients were high enough to provide some evidence of convergent
validity, while the correlation between the two factors was low enough
to provide some evidence of discriminant validity between the
constructs. Additionally, we found that each construct was
differentially related to the various Distinctive Competencies.
Therefore, the results of our analysis provide strong evidence of
validity for our Knowledge Strategy Orientation Scale (KSOS).
By forcing our Explorer and Exploiter scale items onto different
factors our two-factor model resulted in better indices of fit as
evidenced by the superior indices and the Chi-square difference test
over the one-factor model. Our results may be enlightening to
researchers who use employees' perceptions for analysis of
firm-level phenomena. Previous researchers have advocated a self-typing
measure of firm strategy (cf. Hambrick, 1989; Snow and Hambrick, 1980).
These researchers suggest that as an alternative to researcher
inferences about firm strategy from archival data, an
organization's managers might acceptably and reliably characterize
their firm's strategy. Our analysis suggests that perceptions of
firm strategy, as measured by our scale, are consistent regardless of
the respondents' position held in a firm.
Our scales both resulted in acceptably high alpha reliabilities
(Nunnally, 1978). This level of average inter-item correlation suggests
that the Explorer items provide an aggregate measure of one construct
and the Exploiter items measure another separate and distinct construct.
Additionally, our standardized factor pattern coefficients ranged in
Sample Two from .40 to .80 and in Sample Three from .57 to .76 for the
Explorer scale. These values ranged in Sample Two from .39 to .68 and in
Sample Three from .48 to .78 for the Exploiter scale, thus providing
some additional evidence of the convergent validity of our scales.
Regarding discriminant validity, our attenuated correlation between the
two factors ranged from .40 to .55, but our CFA results show a
disattenuated correlation ranging from .59 to .67. Although this
indicates that they are indeed correlated, they are different enough so
as to not be considered collinear (Kline, 1998), indicating some
evidence of discriminant validity. The moderate level of correlation
between our constructs as well as the acceptably high alpha coefficients
indicate that our respondents conceive of the Explorer and Exploiter
knowledge strategy orientations as independent constructs.
Data provided by the measure of our two constructs provided
evidence of differential relationships with the distinctive
competencies. Almost all of our distinctive competencies were
significantly stronger in their relationship with our focal constructs.
Specifically, two of four distinctive competencies associated with
radical innovation were more strongly related to the Explorer knowledge
strategy orientation than to the Exploiter knowledge strategy
orientation. The stronger correlations between Explorer knowledge
strategy orientation and distinctive competencies associated with
radical innovation are consistent with the work of He and Wong (2004).
This is a favorable source of external validation as we expect firms
that engage in radical innovation to utilize an explorer's
knowledge strategy orientation. We also found that each of the
distinctive competencies associated with incremental innovation were
more strongly related to the Exploiter knowledge strategy orientation
than to the Explorer knowledge strategy orientation. These findings are
consistent with the theoretical frameworks provided by Bierly and Daly
(2002), March (1991), and Levinthal and March (1993).
Managerial Implications
One objective of our development of the KSOS was to address some of
the methodological problems of past research in this area: an
over-reliance on archival data, the use of single-item measures, and the
use of proxy measures for focal constructs. Our study focused on a
psychometric evaluation of survey items based upon theoretical insights
provided by Holmqvist (2004), March (1991), Levinthal and March (1993),
Bierly and Chakrabarti (1996), and Zack (1999) regarding firms'
knowledge strategies. Our second objective in development of this new
scale was to provide an instrument that would prove useful to managers
in the areas of organizational assessment and strategic planning.
When compared to measurement based on single-item measures,
archival data, and proxies, a scale such as ours captures the essence of
the Explorer and Exploiter constructs much more accurately. The
relatively simple and easy to use scale allows managers to gather
primary data about their organizations regarding knowledge strategy
orientation, and then use the subsequent analysis to inform decisions
that pertain to knowledge management and innovation in their firms. We
believe the establishment of a commonly accepted measure of knowledge
strategy orientation will help managers to identify organizational
strengths and weaknesses within the focus areas delineated by each of
the KSOS items (e.g., radical versus incremental innovation). Managers
need to be able to objectively evaluate their organization's
knowledge base, discern how knowledge is transferred and integrated in
the organization, and develop knowledge strategies that maximize the
potential of their knowledge base.
Use of the KSOS will help managers to better understand and assess
their strategic choices regarding the creation or acquisition of new
knowledge and the ability to leverage existing knowledge. The
firm's knowledge strategy orientation should help to guide
managerial choices regarding resource allocation in the firm. For
example, an exploration orientation may support the allocation of
additional resources to new product development within a small firm that
relies on advances in technology to ensure its competitive viability. In
low-tech firms resources may best be used to support incremental
continuous improvement and a focus on marketing (e.g., reinforcing brand
image) rather than on attempts at radical innovation. These choices are
particularly important to smaller firms that are more resource
constrained and thus cannot pursue both the Explorer and Exploiter
strategies simultaneously.
Geiger and Cashen (9002) address the important issue of resource
allocation and its affects on innovation in their article exploring the
effects of organizational slack. The KSOS, which helps firms identify
their propensity for radical versus incremental innovation, could be
very useful in future research of these issues. Additionally, the scale
could be used by researchers exploring ways to retain and manage
intellectual capital in organizations. For instance, Droege and
Hoobler's work (2003) discusses the relationship between employee
turn-over and tacit knowledge loss in organizations. Their research
indicates that allocating resources to enhance and promote social
network structures within the organization is vital to retaining tacit
organizational knowledge. Studies such as these indicate the potentially
widespread usefulness of the KSOS measure in further research and
practical application in the management field.
Limitations and Suggestions for Future Research
Scale construction is a dynamic process, with the objective of
continually improving the measurement of a construct (DiTommaso et al.,
2004), and therefore we suggest that the validity characteristics of our
scales need further study. Accordingly, the generalizability of our
findings is limited by the nature of our samples: three (although very
heterogeneous with respect to position held) groups of respondents from
small manufacturing companies located in the mid-Atlantic region of the
U.S. However, we do acknowledge that a conceptual difference exists
between "position" and "level" in an organization
such that some positions exist at different levels and some levels do
not align perfectly across functions. For example, in an accounting
department one might find three levels: accounting directors, cost
accountants, and accounts payable clerks. In a manufacturing department
one might find first-line supervisors, journeyman welders, and welder
helpers. In the organizational hierarchy, accounting directors and
first-line manufacturing supervisors might be on different levels with
only a few levels of management above the former, but numerous levels
above the latter. Clearly, a difference exists between positions,
functions, and levels and future research might seek to explore
differences in perceptions of a firm's knowledge strategy
orientation both between and among positions, functions, and levels in a
particular firm. Another step might be to employ our scales in a
predictive validity study using large organizations in both
manufacturing and service industries.
This would allow for the assessment of the measurement
characteristics of our scales in different samples and in different
domains of industry. We believe that our KSOS can provide researchers
with an alternative measure of firms' strategic orientation and
that employees' perceptions of firm strategy can help overcome some
of the measurement shortcomings of using archival data as proxies for
organizational-level variables.
References
Bentler, P. M. 1990. "Comparative Fit Indexes in Structural
Equation Models." Psychological Bulletin 107: 238-246.
Bergh, D. D. 2001. "Diversification Strategy Research at a
Crossroads." Chapter in Handbook of Strategic Management. Eds. M.
A. Hitt, R. E. Freeman and J. S. Harrison. Oxford, UK: Blackwell. pp.
362-383.
Bierly, P. and A. Chakrabarti. 1996. "Generic Knowledge
Strategies in the U.S. Pharmaceutical Industry." Strategic
Management Journal 17 (Winter Special Issue): 123-135.
--and P. Daly. 2002. "Aligning Human Resource Management
Practices and Knowledge Strategies." Chapter in The Strategic
Management of Intellectual Capital and Organizational Knowledge. Eds. C.
W. Choo and N. Bontis. New York, NY: Oxford University Press. pp.
277-295.
Boyd, B. K., S. Gove and M. A. Hitt. 2005. "Construct
Measurement in Strategic Management Research: Illusion or Reality?"
Strategic Management Journal 26: 239-257.
Browne, M. W. and R. Cudeck. 1993. "Alternative Ways of
Assessing Model Fit." Chapter in Testing Structural Equation
Models. Eds. K. A. Bollen and J. S. Long. Newbury Park, CA: Sage. pp.
445-455.
Comrey, A. L. 1988. "Factor-analytic Methods of Scale
Development in Personality and Clinical Psychology." Journal of
Consulting and Clinical Psychology 5: 754-761.
Damanpour, F. 1991. "Organizational Innovation: A
Meta-analysis of Effects of Determinants and Moderators." Academy
of Management Journal 34: 555-590.
Delaney, J. T. and M. A. Huselid. 1996. "The Impact of Human
Resource Management Practices on Perceptions of Organizational
Performance." Academy of Management Journal 39: 949-969.
Dewar, R. D. and J. E. Dutton. 1986. "The Adoption of Radical
and Incremental Innovations: An Empirical Analysis." Management
Science 32: 1422-1433.
DiTommaso, E., C. Brannen and L. A. Best. 2004. "Measurement
and Validity Characteristics of the Short Version of the Social and
Emotional Loneliness Scale for Adults." Educational and
Psychological Measurement 64: 99-19.
Droege, S. B. and J. M. Hoobler. 2003. "Employee Turnover and
Tacit Knowledge Diffusion: A Network Perspective." Journal of
Managerial Issues 15 (1): 50-64.
Ettlie, J., W. Bridges and R. O'Keefe. 1984.
"Organization Strategy and Structural Differences for Radical
Versus Incremental Innovation." Management Science 30: 682695.
Fan, X., B. Thompson and L. Wang. 1999. "Effects of Sample
Size, Estimation Methods, and Model Specification on Structural Equation
Modeling Fit Indexes." Structural Equation Modeling 6: 56-83.
Floyd, F.J. and K. Widaman. 1995. "Factor Analysis in the
Development and Refinement of Clinical Assessment Instruments."
Psychological Assessment 7: 286-299.
Garcia, R., R. Calantone and R. Levine. 2003. "The Role of
Knowledge in Resource Allocation to Exploration Versus Exploitation in
Technologically Oriented Organizations." Decision Sciences 34:
323-349.
Geiger, S. W. and L. H. Cashen. 2002. "A Multidimensional
Examination of Slack and Its Impact on Innovation." Journal of
Managerial Issues 14 (1): 68-85.
Godfrey, P. C. and C. W. L. Hill. 1995. "The Problem of
Unobservables in Strategic