International Journal of Information Management 29 (2009) 210–218

Contents lists available at ScienceDirect

International Journal of Information Management

journal homepage: www.elsevier.com/locate/ijinfomgt

Social capital, behavioural control, and tacit knowledge

sharing—A multi-informant design
Shu-Chen Yang a,∗ , Cheng-Kiang Farn b
a
Department of Information Management, National University of Kaohsiung, No. 700 Kaohsiung University Road, Nan Tzu District, Kaohsiung, Taiwan
b
Department of Information Management, National Central University, Taiwan

a r t i c l e i n f o a b s t r a c t

Keywords: As suggested by prior studies, tacit knowledge sharing is a natural process of social interaction. The
Behavioural control perspectives of social capital and behavioural control are thus employed in this study to investigate an
Intention–behaviour relationship employee’s tacit knowledge sharing and behaviour within a workgroup. This study collects data through a
Tacit knowledge sharing
multi-informant questionnaire design. Three interesting results were obtained in this study. First, results
Social capital
show that tacit knowledge sharing intention can be induced by affect-based trust. However, shared value
is negatively related to tacit knowledge sharing intention. Second, internal control has a positive effect
on tacit knowledge sharing intention, but the relationship between internal control and tacit knowledge
sharing behaviour could not be confirmed. Third, external control positively moderates the relationship
between tacit knowledge sharing intention and behaviour. It is interesting to note that tacit knowledge
sharing intention does not necessarily lead to tacit knowledge sharing behaviour unless the moderating
effect of external control is taken into account. These findings and their implications are also addressed.
© 2008 Elsevier Ltd. All rights reserved.

1. Introduction
The importance of knowledge within organisations has been
highlighted by several researchers (e.g. Alavi & Leidner, 2001; Bock,
Zmud, Kim, & Lee, 2005). It is important to note that knowledge
itself cannot create significant value without utilisation (Fahey
& Prusak, 1998). As argued by Alavi (2000), knowledge sharing
among organisational members is the most important and chal-
lenging means to increase the value of knowledge utilisation.
Based on Polanyi (1967) conceptualisation, Nonaka (1994) sug-
gested that knowledge can be classified as tacit and explicit. Tacit
knowledge—reflecting an individual’s know-how and experiences
from past actions—is increasingly considered as a valuable intan-
gible resource that is difficult to imitate and acquire, and can be
regarded as the most important source of competitive advantage for
an individual, a group, or a firm (Berman, Down, & Hill, 2002). This
is especially true in the context of innovative works, where much of
the task-related knowledge is tacit in nature, and tacit knowledge
sharing among members is crucial for creating higher collective
performance (Käser & Miles, 2002). However, an individual may
hoard rather than share his/her tacit knowledge because it is valu-
∗ Corresponding author. Tel.: +886 7 5919326; fax: +886 7 5919328.
E-mail address: henryyang@nuk.edu.tw (S.-C. Yang).
able and important, and the contribution of tacit knowledge cannot
be easily measured and compensated accordingly (Osterloh & Frey,
2000). Thus, tacit knowledge acquisition and sharing is one of most
important issues for knowledge management within organisations.
For a workgroup, tacit knowledge sharing among members is also
critical for task completion and group performance. Accordingly,
this study intends to answer the following two research questions:
(1) what are the factors that influence an individual’s tacit knowl-
edge sharing intention and behaviour within a workgroup? (2) Will
an individual’s tacit knowledge sharing intention necessarily lead
to tacit knowledge sharing behaviour?
While explicit knowledge sharing can be facilitated by infor-
mation technology, tacit knowledge sharing is subject to social
interaction (Käser & Miles, 2002; Nonaka, 1994). In other words,
tacit knowledge sharing among organisational members is socially
driven. Moreover, knowledge sharing behaviour is inherently a type
of collective action (Bock et al., 2005) and sometimes beyond an
individual’s volitional control. Accordingly, perspectives of social
capital and behavioural control are employed in this study in order
to investigate an employee’s tacit knowledge sharing behaviour
within a workgroup. In this study, data are collected through multi-
informant questionnaires in which each respondent’s social capital
and tacit knowledge sharing behaviour are reported by his/her
colleagues. In this manner the study avoids the perceptual bias
from self-reporting and common-method bias from single infor-
mants.

0268-4012/$ – see front matter © 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijinfomgt.2008.09.002
 S.-C. Yang, C.-K. Farn / International Journal of Information Management 29 (2009) 210–218
2. Conceptual background
2.1. Tacit knowledge sharing
Nonaka (1994) addresses two types of knowledge within organ-
isations: explicit and tacit. Explicit knowledge is regarded as
knowledge that can be formally and systematically stored, artic-
ulated, and disseminated in certain codified forms such as manual
or computer files (Becerra-Fernandez & Sabherwal, 2001). On the
other hand, tacit knowledge is deeply rooted in action, experience,
thought, and involvement in a particular context (Alavi & Leidner,
2001), and thus is difficult to be transformed into explicit form in
order to be easily transferred and shared (Berman et al., 2002).
Polanyi (1967, p. 4) describes the nature of tacit knowledge with
the following phrase: ‘we know more than we can tell’. That is to
say, tacit knowledge is deeply embedded in the mind to the extent
that the knowers are not fully aware of that knowledge they pos-
sess (Koskinen, Pihlanto, & Vanharanta, 2003). Nevertheless the
tacit knowledge determines the behaviour of the knower. Common
examples of tacit knowledge include the ability to ride a bicycle,
the knowledge of an expert baseball player, and skill debugging a
computer program.
Tacit knowledge may be understood through the concepts of
skill (Berman et al., 2002) or practical know-how (Koskinen et al.,
2003), thus an individual usually will not share his/her knowl-
edge when the knowledge is regarded as valuable or important
because of a fear of losing possible advantages (Bock et al., 2005).
Furthermore, potential risk of losing advantage and lack of proper
reward mechanism are the major reasons that an individual is usu-
ally reluctant to share his/her tacit knowledge with others (Osterloh
& Frey, 2000). Thus, tacit knowledge sharing can be only facilitated
by intrinsic motivation, such as sociability and friendship (Osterloh
& Frey, 2000). Choi and Lee (2003) also suggest that an individual
can acquire tacit knowledge and personal experience only in a tacit-
oriented manner that emphasises social interaction. Nonaka (1994)
also considers that tacit knowledge is of a personal nature and
can be shared through sharing metaphors or experiences during
social interaction without substantial knowledge loss. Thus, social
relationship may be the most important factor that facilitates tacit
knowledge sharing among employees within an organisation.
2.2. Social capital
Social capital is conceptualised as a set of resources embedded
in the social relationship among social actors and can be regarded
as a valuable asset that secures benefits for social actors ranging
from individuals to organisations (Adler & Kwon, 2002). For exam-
ple, a higher degree of social capital is helpful in finding better jobs
(Granovetter, 1995), receiving early promotions (Burt, 1997), mak-
ing collective work easier (Bolino, Turnley, & Bloodgood, 2002),
inter-unit resource exchange and collaboration within an organi-
sation (Tsai & Ghoshal, 1998), creation of intellectual capital and
dissemination of knowledge within an organisation (Nahapiet &
Ghoshal, 1998), enhancing organisational flexibility (Leana & Van
Buren, 1999), and facilitating IT outsourcing (Chou, Chen, & Pan,
2006).
In a broader sense, social capital is not a unidimensional con-
cept (Putnam, 1995) and ‘encompasses many aspects of a social
context, such as social ties, trusting relations, and value systems
that facilitate actions of individuals located within that context’
(Tsai & Ghoshal, 1998, p. 465). Putnam (1995) argues that clar-
ifying the dimensions of social capital is a top priority because
social capital has many complicated attributes related to a social
context. Accordingly, Nahapiet and Ghoshal (1998) propose that
structural, relational, and cognitive aspects are the three dimen-
211
sions of social capital. This three-dimensional framework has been
employed to investigate the relationship between social capital and
intra-organisational phenomena, such as creation of intellectual
capital (Nahapiet & Ghoshal, 1998), inter-unit resource exchange
(Tsai & Ghoshal, 1998), and organisational citizenship behaviour
(Bolino et al., 2002).
Structural social capital can be conceptualised as the overall
pattern of relationships among social actors (Nahapiet & Ghoshal,
1998). Bolino et al. (2002) suggest that the structural social capi-
tal can also be considered as the extent to which actors in a social
network are connected. Relational social capital includes the assets
created and leveraged through ongoing relationship that influence
social actors’ behaviour (Nahapiet & Ghoshal, 1998). This dimen-
sion bears some resemblance to Adler and Kwon’s (2002) concept
of ‘goodwill’ and can be manifested by trust, norms, obligations,
and identification (Nahapiet & Ghoshal, 1998). Besides, Nahapiet
and Ghoshal (1998) regard cognitive social capital as the com-
mon understanding among social actors through shared language
and narratives. It is embodied in attributes such as shared vision
or shared value that facilitates individual and collective actions
and common understanding of proper actions and collective goals.
Boland and Tenkasi (1995) suggest that higher cognitive social
capital gives partners a common perspective that enables them
to develop similar perception and interpretation toward events.
Unlike the impersonal nature of structural social capital, both
relational and cognitive dimensions describe the personal qual-
ities of interpersonal relationship (Bolino et al., 2002) and can
be categorised into relational embeddedness of social capital that
represents the motivational characteristic of interpersonal social
exchange. Thus, relational social capital and cognitive social capital
are emphasised in this study.
Among the studies on relational social capital, trust quality
has received much attention in organisational and management
research (McEvily, Perrone, & Zaheer, 2003). Further, trust is an
important dimension of Leana and Van Buren’s (1999) conceptu-
alisation of social capital. McAllister (1995) suggests that the two
factors of trust are ‘affect-based trust’ and ‘cognition-based trust’.
Trust is cognition-based in that ‘we cognitively choose whom we
will trust in which respects and under which circumstances, and we
base the choice on what we take to be “good reasons”, constitut-
ing evidence of trustworthiness’ (Lewis & Weigert, 1985, p. 970).
On the other hand, affect-based trust is based on the emotional
ties linking individuals, such as friendship, love, or care (Lewis &
Weigert, 1985; McAllister, 1995). This study employs affect-based
trust to characterise the relational dimension because relational
social capital represents the affective quality of interpersonal rela-
tionship (Bolino et al., 2002). Furthermore, the concept of shared
value is employed to characterise the cognitive dimension of social
capital in this study. Morgan and Hunt (1994, p. 25) define shared
value as ‘the extent to which partners have beliefs in common about
what behaviours, goals, and polices are important or unimportant,
appropriate or inappropriate, and right or wrong’.
2.3. Internal and external control
Although there are several empirical studies that support the
Theory of Planned Behaviour (TPB) across a number of domains (e.g.
Armitage & Conner, 2001; Conner & McMillan, 1999), the role and
concept of perceived behavioural control (PBC) has not yet garnered
a consensus. In the earlier versions of the TPB, Ajzen (1985) states
that PBC, which reflects the availability of requisite opportunities
and resources, will moderate the intention–behaviour relationship.
That is to say, the relationship between intention and behaviour is
stronger when PBC is high. Due to lack of evidence on the moderat-
ing effects of PBC, Ajzen (1991) subsequently argues that PBC will
212 S.-C. Yang, C.-K. Farn / International Journal of Information Management 29 (2009) 210–218
affect behaviour directly. However, the moderating effect of PBC
has been supported by several empirical studies (e.g. Armitage &
Conner, 2001; Conner & McMillan, 1999; Terry & O’Leary, 1995).
Conner and Armitage (1998) suggest that PBC can be dimension-
alised into internal and external control. Internal control includes
intrinsic control factors and is similar to the concept of self-efficacy,
whereas external control reflects extrinsic control factors and is
similar to Triandis’s (1977) concept of facilitating conditions or Kuhl
and Beckmann’s (1985) ‘action control’. The distinctions between
internal control and external control also embody Ajzen’s (1985)
suggestion that non-motivational factors may relate to both inter-
nal aspects (such as skill or ability) and external aspects (such as
opportunities or resources). The external control factors can also
be described in terms of available resources and opportunities that
will facilitate the performance of focal behaviour (Fitch & Ravlin,
2005). Ajzen (2002) suggested that PBC can lead to distinct com-
ponents of self-efficacy and controllability, which are similar to
internal control and external control, respectively.
3. Research model and hypotheses development
Based on the previous conceptual background, this study pro-
poses a research model in order to investigate tacit knowledge
sharing intention and behaviour within a workgroup (as shown in
Fig. 1).
Social capital can reflect strong interpersonal connections and
extensive investment in interpersonal relationship. As an individ-
ual has higher social capital in his/her social network, he/she will
be likely to behave in ways that benefit other members in order to
maintain an interpersonal friendship in their social network. There-
fore, an individual will be willing to share his/her tacit knowledge
with co-workers through the reciprocal relationship when he/she
has a stronger social capital. This leads to the following hypothesis:
Hypothesis 1a. Affect-based trust positively affects tacit knowl-
edge sharing intention.
Hypothesis 1b. Shared value positively affects tacit knowledge
sharing intention.
As argued by TPB model, behavioural intention captures all the
motivational factors that reflect the effort people plan to exert in
order to perform the behaviour in question (Ajzen, 1991). Thus,
when an individual expresses stronger intention to engage in a cer-
tain type of behaviour, he/she is more likely to put the behaviour
into practice. The following hypothesis is then proposed:
Hypothesis 2. Tacit knowledge sharing intention positively affects
tacit knowledge sharing behaviour.
The TPB model suggests that an individual’s perception of
non-motivational factors (i.e. PBC) plays an important role in the
intention–behaviour relationship when the focal behaviour is not
under complete volitional control (Ajzen, 1991). However, the
effects of PBC have not received consensus among different studies.
On the one hand, as PBC is conceptualised as an internal non-
motivational factor (such as self-efficacy), it is suggested to directly
affect behavioural intention (e.g. Ajzen, 1991; Terry & O’Leary,
1995). PBC is conceptualised as an external non-motivational fac-
tor; on the other hand, it is said to directly affect behaviour or to
moderate the intention–behaviour relationship (e.g. Fitch & Ravlin,
2005; Terry & O’Leary, 1995). In order to clarify the role of PBC in an
individual’s tacit knowledge sharing behaviour, this study adopts
Conner and Armitage’s (1998) dimensionalisation of PBC as internal
control and external control.
According to Conner and Armitage’s (1998) definition, inter-
nal control refers to an individual’s confidence to perform the
behaviour in question. An individual with high internal con-
trol may believe that he/she is able to easily handle the focal
behaviour by him/herself, thus leading to a strong behavioural
intention. However, external non-motivational factors (external
control) exert their moderating effects on the intention–behaviour
relationship when the focal behaviour is under partial volitional
control (Armitage & Conner, 2001). On the other hand, external
control describes all the environmental factors that may impede
or facilitate focal behaviour performance and usually cannot be
manipulated by an individual alone. Higher external control will
facilitate the implementation of intentions into action. In gen-
eral, an individual’s performance in knowledge sharing behaviour
is sometimes beyond his/her own control (not under complete
volitional control). For example, an individual who has strong moti-
vation to share his/her tacit knowledge may not actually provide
his/her know-how and experiences when other members do not
request it. Besides, unimpeded and free communication among
employees is also one of the most important facilitating factors for
tacit knowledge sharing behaviour within an organisation. Accord-
ingly, the relationship between an individual’s tacit knowledge
sharing intention and behaviour will be stronger when he/she per-
ceives that the external conditions are favourable for engaging in
tacit knowledge sharing behaviour. Thus, the following hypotheses
regarding the effects of internal and external control are proposed:
Hypothesis 3. Internal control positively affects tacit knowledge
sharing intention.
Hypothesis 4. Internal control positively affects tacit knowledge
sharing behaviour.
Hypothesis 5. The effect of tacit knowledge sharing intention on
tacit knowledge sharing behaviour is stronger for higher external
control than for lower external control.
4. Research method
4.1. Procedure
This study targeted individuals who are engaged in jobs that
are highly knowledge-intensive and involve certain degree of inter-
personal interaction for task completion. These individuals include
personnel in MIS departments, R& D departments, or various
project teams. In such groups, lateral knowledge sharing among co-
workers occurs more frequently than knowledge sharing between
subordinates and managers. Besides, lateral knowledge sharing
among co-workers is more susceptible to social relationship,
whereas knowledge sharing between subordinates and managers
is usually limited to a formal relationship. Thus, managers in such
groups were not considered as our target respondents. Each depart-
ment or team that satisfies above requirements is referred to as a
qualified group hereinafter.
This study employed a multi-informant design to measure each
respondent’s social capital and tacit knowledge sharing behaviour.
The use of a single informant for all variables in one study is
often criticised for the perceptual bias and common-method bias
involved. In our study, an individual’s social capital and tacit knowl-
edge sharing behaviour could be observed by other members in the
same workgroup. An individual’s social capital refers to the degree
of trust he/she earned and his/her shared values with the other
members in the workgroup. Thus, it is reasonable to measure an
individual’s social capital on the basis of his/her co-workers’ view-
points. Besides, we also measured an individual’s tacit knowledge
sharing behaviour based on the viewpoints of his/her co-workers
in order to eliminate the common-method bias that results when
both the intention and the behaviour were evaluated by oneself. The
 S.-C. Yang, C.-K. Farn / International Journal of Information Management 29 (2009) 210–218
Fig. 1. Research model.
perceptual bias from self-reporting and the common-method bias
from single informants in one study can be overcome when these
two constructs are evaluated by two peers. Multi-informant design
has also been adopted by several studies to address the above prob-
lems (e.g. Hogel, Weinkauf, & Gemuenden, 2004; Simons, Pelled, &
Smith, 1999).
In order to reduce respondents’ tasks to a more manageable
size, three members were invited from each qualified group. Each
respondent was asked to rate actual tacit knowledge sharing
behaviour and social capital for each of the other two members in
his/her qualified group. For example, let us consider R1, R2, and R3
as the three respondents in a qualified group (as shown in Table 1).
R1 was asked to evaluate the social capital and tacit knowledge
sharing behaviour of R2 and R3, while his/her social capital and
tacit knowledge sharing behaviour were evaluated by R2 and R3.
All other remaining constructs were measured through self-ratings
way by each respondent.
This study follows three steps in the data collection. First,
because of the complexities of our research design, we selected
qualified groups from several organisations with which the authors
were in frequent contact in order to ensure a high participation
rate in our sampling process. From each qualified group, three
randomly selected members were asked to participate in this
study. The random selection process continued for each qualified
group until three members volunteered to participate. For each
respondent, he/she was asked to report not only the other two
members’ social capital and tacit knowledge sharing behaviour
but also other self-reporting constructs. In order to ensure sim-
plicity in our questionnaire design, questionnaires for these three
respondents of a qualified group were tailored to meet research
design. That is to say, in each respondent’s questionnaire, two other
members’ names were explicitly mentioned in the measurement
items for social capital and tacit knowledge sharing behaviour. For
instance, in the above example of a qualified group with R1, R2,
and R3, the names of R2 and R3 were mentioned in R1’s question-
naire.
Second, the tailored questionnaires were then sent by e-mail.
Each respondent was asked to fill out his/her tailored question-
naire in a computer file and reply by e-mail after completion
for the sake of confidentiality. We also sent follow-up e-mails to
those who did not return their finished questionnaires within 3
days. Third, as each respondent’s social capital and tacit knowl-
edge sharing behaviour were reported by the two other members,
all members in a qualified group were required to complete and
213
return their questionnaires for further analyses. That is to say,
if any one member failed fill out his/her questionnaire, the data
pertaining to his/her fellow members in the same group were
discarded. In order to encourage all three members in a qual-
ified group to participate in our study, we sent NTD 200 (in
New Taiwan Dollars) as an incentive to each respondent after all
respondents in the same group completed their tailored question-
naires.
4.2. Measurement
All constructs were measured using multiple-item scales, drawn
from pre-validated measures in the previous related studies (see
Appendix A). In order to ensure the accuracy of the translation
from English to Chinese, this study involved the following three
steps. First, all the items were translated into Chinese with mod-
ification by the authors, one MIS professor, and four MIS doctoral
students who had an average of 5 years of actual experience in sys-
tem development. Second, one graduate student (hereafter referred
to as SG) and one senior undergraduate student (SU) from the
department of English in a University from Northern Taiwan were
asked to translate the Chinese items into English. In order to check
the semantic differences, the two back-translated English versions
from SG and SU were compared with the original English ver-
sion by SU and SG, respectively. The translated Chinese items that
had significant semantic differences were then modified. Third,
a pre-test with several respondents who had real-world work
experience was conducted to ensure the translation was coher-
ent.
5. Analyses and results
Three hundred and six respondents from 67 organisations
agreed to participate in this study. Most respondents are employed
in the Computer and Electronics, Software and Information System,
and Consulting Industry. These respondents were from 102 groups
that ranged in size from 3 to 26 people, with an average team size
of 6.30 (S.D. = 3.67). Sixty percent of the respondents were male
and 76% were 26–34 years old. With regard to tenure, the respon-
dents’ work experience ranged from 1 month to 290 months with
an average of 29.68 months (S.D. = 29.92). Half of these respondents
are employed in R&D department. The characteristics of these 67
organisations and 102 groups are shown in Table 2.
214 S.-C. Yang, C.-K. Farn / International Journal of Information Management 29 (2009) 210–218

Table 1
An example of research design in a qualified group.

Note: R1, R2, and R3 are three respondents in a qualified group. The value in each shadow cell represents the
target(s) to be evaluated.

5.1. Convergent and discriminant validity can be usually detected by inspecting of univariate distributions.
As shown in Table 3, the absolute values of skewness and kurto-
In this study, an individual’s social capital and tacit knowl- sis are smaller than 3.0 and 8.0, respectively, which are all well
edge sharing behaviour were evaluated by two other members in above the threshold suggested by Kline (2005). Thus, the multi-
his/her workgroup. The two responses for each of these two con- variate normality of our data was assured. Besides, the number of
structs are then summed up into a single score of each respondent’s observations (18 observed variables lead to 18 × 17 = 153 observa-
social capital and tacit knowledge sharing behaviour. According to tions) is greater than the number of free estimated parameters (18
the two-step procedure recommended by Anderson and Gerbing observed variables and 5 unobserved variables lead to 46 free esti-
(1988), we estimated and re-specified the measurement model mated parameters) in this study; thus, the CFA model in our study
prior to examining the structural model. A CFA was used to assess is over-identified for further analyses (Kline, 2005).
the convergent and discriminant validity of the operationalisa- We first performed the CFA with 18 items (see Appendix B for the
tion. All the research constructs were modelled as five correlated covariance matrix) and the threshold employed for judging the sig-
first-order factors and employ LISREL 8.70 using the Maximum nificance of factor loadings was 0.60 suggested by Sharma (1996).
Likelihood (ML) estimation to estimate the measurement model. As the measurement model estimation shown in Table 3, factor
The construct external control was not included in the CFA analy- loadings of all measurement items range from 0.712 to 0.967, which
sis because it is measured with formative indicators. As suggested indicates acceptable convergent validity. The CFA resulted in a chi-
by Cohen, Cohen, Teresi, Marchi, and Velez (1990), the reliability square statistic of 339.396 with 125 degrees of freedom. Since the
criteria employed for characterising the constructs measured by chi-square is less than three times the degrees of freedom, a good
reflective indicators cannot be applied to the construct measured fit is implied (Carmines & McIver, 1981). Furthermore, the values on
by formative indicators. other goodness of fit indexes also show a relative good fit between
Based on the basic assumptions of the CFA approach proposed by the measurement model and data.
Kaplan (2000), the independence assumption was met due to our This study also assessed construct reliability by calculating com-
random sampling design and there are also no missing values in our posite reliability that assesses whether the specified indicators are
dataset. Kline (2005) suggested that multivariate non-normality sufficiently representative of their respective latent factors, as sug-
gested by Segars (1997). These estimates of composite reliability
Table 2
of latent factors range from 0.91 to 0.96, which are all well above
Demographic characteristics of sample (n = 306).
the threshold of 0.70 suggested by Jöreskog and Sörbom (1989);
Characteristics Percentage thus, acceptable construct reliability is implied. However, compos-
Industry Banking 4.9 ite reliability does not reflect the degree of variance that is captured
Computer and 28.4 by the construct in relation to the amount of variance due to mea-
Electronics
surement error (Fornell & Larcker, 1981). Thus, average variance
Consulting 21.6
Government 2.9
extracted (AVE) estimate was employed to obtain this information.
Logistics 5.9 AVE estimate of 0.50 or higher indicates acceptable validity for a
Mobile Phone 6.9 construct’s measure (Fornell & Larcker, 1981). All AVE estimates in
Software and 24.5 this study (range from 0.72 to 0.90) are well above the cut-off value,
Information System
which suggests that all measurement scales have convergent valid-
Others 4.9
ity. In order to assess discriminant validity among the constructs,
Group type MIS Department 23.5 we calculated square root of AVE for each construct and compared
Project Team 26.5
R&D Department 50.0
them with inter-construct correlations for each pair of constructs.
Result also shows that square roots of all AVE estimates for each
Group size (average = 6.30, Less than 5 51.0
construct are greater than inter-construct correlations; thus, dis-
S.D. = 3.67) 6–9 35.3
Over 10 13.7 criminant validity is supported.

Gender Female 39.9

Male 60.1 5.2. Direct effect testing
Age Under 25 7.8
26–29 37.6 In order to test the direct effect hypotheses in our research
30–34 37.9 model, a structural equation modelling (SEM) analysis was per-
35–39 12.1
formed. We modelled six first-order factors (all constructs in this
40–49 4.6
Over 50 0.0 study except for external control) with five hypothetical causal
paths and employ LISREL 8.70 with the Maximum Likelihood esti-
Tenure (average = 29.68, Less than 12 months 27.8
S.D. = 29.92) 13–24 months 33.6
mation to estimate the structural model. Fig. 2 shows that the
25–36 months 13.8 structural model estimation resulted in a chi-square statistic of
37–60 months 13.4 318.284 with 123 degrees of freedom. The chi-square is less than
61–120 months 10.4 three times the degrees of freedom; thus, a good fit is implied
Over 121 months 1.0
(Carmines & McIver, 1981). Furthermore, the values on other good-
 S.-C. Yang, C.-K. Farn / International Journal of Information Management 29 (2009) 210–218 215

Table 3
Measurement model estimation and basic statistics.

Construct Item Mean S.D. Item loading t-Value Error variance Skewness Kurtosis

Affect-based trust ABT1 11.255 1.982 0.814 17.095 0.338 −0.063 −0.007
ABT 2 11.258 2.118 0.912 20.549 0.169 −0.072 −0.043
ABT 3 9.944 2.514 0.812 17.048 0.340 −0.042 −0.040
ABT 4 10.729 2.214 0.913 20.598 0.167 −0.047 −0.038
ABT 5 10.477 2.330 0.874 19.134 0.236 −0.042 −0.044

Shared value SHV1 9.680 2.100 0.861 18.543 0.259 −0.014 0.017
SHV2 10.507 2.044 0.901 19.982 0.188 −0.037 0.005
SHV3 10.771 2.053 0.896 19.795 0.197 −0.046 0.000
SHV4 9.624 2.097 0.712 14.011 0.493 −0.009 0.012

Tacit knowledge TKSI1 6.229 0.958 0.956 22.334 0.086 −0.635 −0.472
sharing intention TKSI2 6.294 0.930 0.924 21.056 0.145 −0.698 −0.389
TKSI3 6.170 0.957 0.905 20.308 0.181 −0.566 −0.551

Tacit knowledge TKSB1 11.000 2.170 0.944 22.002 0.108 −0.060 −0.074
sharing behaviour TKSB2 11.402 2.096 0.960 22.683 0.078 −0.081 −0.052
TKSB3 10.974 2.124 0.939 21.775 0.118 −0.058 −0.061

Internal control ITC1 5.314 1.228 0.912 20.356 0.168 −0.177 −0.255
ITC2 5.438 1.208 0.967 22.479 0.066 −0.208 −0.257
ITC3 5.614 1.185 0.814 17.083 0.337 −0.273 −0.424

2 (d.f.) = 339.396 (125); RMSEA = 0.0750; CFI = 0.979; NFI = 0.966; NNFI = 0.974; GFI = 0.890.

Fig. 2. Structural model estimation.

ness of fit indexes also show a relative good fit between structural
model and data.
As shown in Fig. 2, two out of five hypothetical causal paths
are not significant. First, the effect of affect-based trust on tacit
knowledge sharing intention is the opposite of that of shared value.
Second, internal control positively affects tacit knowledge sharing
behaviour while the relationship between internal control and tacit
knowledge sharing behaviour is not significant. Third, results do not
strongly support the direct effect of tacit knowledge sharing inten-
tion on tacit knowledge sharing behaviour. That is, an individual
may not put his/her strong tacit knowledge sharing intention into
action; this indicates the importance of situational factors.
5.3. Moderating effect testing
Model 2 : TKSB = b0 + b1 TKSI + b2 ITC + b3 ETC
Model 3 : TKSB = b0 + b1 TKSI + b2 ITC + b3 ETC + b4 ETC × TKSI
Table 4 shows the results of MRA for the moderating effect of
external control on the relationship between tacit knowledge shar-
ing intention and behaviour. This study found significant change in
R2 between Model 2 and Model 3, which indicates the significant
moderating effect of external control. The relationship between
tacit knowledge sharing intention and tacit knowledge sharing
behaviour is stronger under higher external control than under
lower external control. Thus, Hypothesis 5 is supported. This study
also found that the change in R2 from Model 1 to Model 2 is signif-
icant; external control is thus regarded as a quasi-moderator.

This study employed moderated regression analyses (MRA) to Table 4

MRA for moderating effect of ETC on TKSI → TKSB.
identify the moderating effects without information loss result-
ing from the artificial transformation of a continuous variable into Models R2 R2 F for R2
a qualitative one in the subgroup analyses (Szymanski, Troy, & 1 TKSB = b0 + b1 TKSI + b2 ITC 0.016 – 2.444
Bharadwaj, 1995). Three models in MRA for moderating effect of 2 TKSB = b0 + b1 TKSI + b2 ITC + b3 ETC 0.471 0.455 259.459***
external control are designed as follows: 3 TKSB = b0 + b1 TKSI + b2 ITC + b3 ETC + b4 ETC × TKSI 0.478 0.007 4.208*
*
p < 0.05.
Model 1 : TKSB = b0 + b1 TKSI + b2 ITC ***
p < 0.001.
216 S.-C. Yang, C.-K. Farn / International Journal of Information Management 29 (2009) 210–218
6. Discussion and implications
As suggested by the previous literature, tacit knowledge sharing
is subject to social interaction (e.g. Käser & Miles, 2002; Nonaka,
1994; Osterloh & Frey, 2000); this study investigates tacit knowl-
edge sharing among workgroup members from the perspective of
social capital and behavioural control. Results can be employed
to answer the two research questions described in Section 1.
First, an employee’s tacit knowledge sharing intention is affected
by affect-based trust, shared value, and internal control. Second,
an employee’s tacit knowledge sharing intention does not nec-
essarily lead to tacit knowledge sharing behaviour. However, the
external control for knowledge sharing moderates the relation-
ship between an employee’s tacit knowledge sharing intention and
behaviour. Furthermore, it is interesting to note that social capi-
tal may be a double-edged sword, which can both facilitate and
hinder knowledge sharing within a workgroup. As suggested by
our results, shared value is negatively related to tacit knowledge
sharing intention. One possible interpretation is that Taiwanese’s
cultural values discourage employees to share knowledge, as men-
tioned in Hsu’s (2006) study. Most employees in Taiwan are not
accustomed to sharing experiences and know-how unless they
have an intimate relationship with other colleagues (as suggested
in Hypothesis 1). When an employee holds common understand-
ing with other members who usually hoard knowledge in the same
workgroup, he/she may also be reluctant to share his/her tacit
knowledge.
Future research may be conducted with the limitations of this
study in mind. First, three members were selected as respondents
for each qualified group in order to reduce respondents’ tasks
to a more manageable size. However, this research design may
not be suitable for a workgroup with a large number of mem-
bers. Second, as this study employed a cross-sectional design,
all the hypothetical causal relationships can only be inferred
rather than proven. A study with longitudinal design can enhance
our understanding about the dynamics of tacit knowledge shar-
ing among employees. Third, our findings were derived based
on the sample from MIS departments, R&D departments, and
project teams. However, one needs to be cautious while apply-
ing our findings to other kinds of groups. Finally, the cultural
factors should be taken into account while interpreting our
results because this study was conducted in Taiwan. People’s atti-
tudes and tendency to share knowledge in corporations in the
East may be quite different from those in corporations in the
West.
There are several theoretical implications of our study. First,
this study employed perspectives of social capital to investigate
an employee’s tacit knowledge sharing intention and behaviour
within his/her workgroup. Several literatures have claimed that
tacit knowledge sharing among employees is social driven, but
extant empirical studies about antecedents of employees’ tacit
knowledge sharing are not abundantly available. This study pro-
vides a compelling theoretical framework for conducting an
empirical study for this line of research. Future studies can extend
this study to better explicate tacit knowledge sharing within
organisations. Second, this study investigated both employee’s
tacit knowledge sharing intention and behaviour in an inte-
grated framework. Most knowledge sharing studies did not take
the actual behaviour into account by assuming that behavioural
intention is highly related to behaviour, as suggested by TPB mod-
els (e.g. Liao, Shao, Wang, & Chen, 1999). However, this study
found that an employee’s tacit knowledge sharing intention is
not significantly related to tacit knowledge sharing behaviour
and the intention–behaviour relation is positively moderated by
his/her external control about tacit knowledge sharing. This study
performed a rather insightful examination of employees’ tacit
knowledge sharing. Finally, a multi-informant research design was
employed in this study. It is reasonable that an employee’s social
capital and tacit knowledge sharing behaviour are reported by
his/her colleagues in the same workgroup. Such a design can avoid
the perceptual bias from self-reporting and the common-method
bias from single informants and is useful for conducting a solid
empirical study.
The empirical results of this study also have several manage-
rial implications. First, the results show that affect-based trust is
an important prerequisite for effective interpersonal tacit knowl-
edge sharing. Thus, managers need to foster the formation of
an intensive social network among employees in order to pro-
mote tacit knowledge sharing within a workgroup through their
exhibition of extra-role behaviours. Besides, employees in Taiwan
are usually reluctant to share knowledge due to an authoritarian
mindset and a desire to protect prestige (Hsu, 2006). Managers
should aim to cultivate a sharing environment to alleviate this cul-
tural burden. Second, internal control is an important determinant
of employees’ tacit knowledge sharing intention. An employee’s
internal control about tacit knowledge sharing usually derives
from his/her individual characteristics and organisational experi-
ences. When an individual is usually encouraged to share his/her
tacit knowledge and encounters fewer obstacles, he/she will be
confident of his/her ability to share tacit knowledge with col-
leagues.
Third, external control has a contingent effect on an employee’s
tacit knowledge sharing. When an employee perceives that he/she
has no opportunities for tacit knowledge sharing, he/she will not
actually share his/her experiences or know-how with others even
though he/she has strong willingness to do so. For managers,
this result implies that offering free communication mechanisms
that are favourable for tacit knowledge sharing among employ-
ees is of paramount importance. When an employee can be well
aware of the problems and difficulties encountered by other col-
leagues and the communication channels among employees are
unhindered, he/she will easily put his/her tacit knowledge shar-
ing intention into action. For example, a group meeting can be
held frequently to offer opportunities for tacit knowledge sharing
among employees. Thus, it is very important for managers to reduce
the obstacles and difficulties in tacit knowledge sharing among
employees.
7. Conclusion
This study used responses from 306 employees in 102 work-
groups across 67 organisations in order to examine the roles that
social capital and behavioural control play in tacit knowledge shar-
ing and behaviour among organisational members. The results
reveal that social capital may be a double-edged sword for tacit
knowledge sharing intention and that external control plays a
moderating role in intention–behaviour relationship. Despite its
limitations, we believe that this study may be useful for future
research on knowledge sharing.
Appendix A. Measurement items
Affect-Based Trust (adapted from McAllister, 1995)
1. We have a sharing relationship. We can both freely share our
ideas, feelings, and hopes.
2. I can talk freely to this individual about difficulties I am having
at work and know that he/she will want to listen.
 S.-C. Yang, C.-K. Farn / International Journal of Information Management 29 (2009) 210–218 217

3. We would both feel a sense of loss if one of us was transferred
and we could no longer work together in the same group.
4. If I shared my problems with this person, I know he/she would
respond constructively and caringly.
5. I would have to say that we have both made considerable emo-
tional investments in our personal relationship.
Shared Values (adapted from Brashear, Boles, Bellenger, & Brooks,
2003)
Tacit Knowledge Sharing Behavior (adapted from Bock et al., 2005)
1. He/she always shares his/her working experience or know-how
with other group members.
2. He/she always shares his/her ways to solve problems at the
request of other group members.
3. He/she always tries to share his/her expertise from his/her edu-
cation or training with other group members in a more effective
way.

Internal Control (Armitage, Conner, Loach, & Willetts, 1999)

1. I feel that my personal values are a good fit with those of he/she.
2. He/she has the same values as I have with regard to the distri-
bution of work within our group.
3. He/she has the same values as I have with regard to the purpose
of our group.
4. In general, my values and the values held by he/she are very
similar.
Tacit Knowledge Sharing Intention (adapted from Bock et al., 2005)
1. I intend to share my working experience or know-how with other
group members more frequently.
2. I am willing to share my ways to solve problems at the request
of other group members.
3. I am willing to try to share my expertise from my education
or training with other group members in a more effective
way.
1. I believe I have the ability to share my working experience or
know-how with other group members.
2. I am confident to share my working experience or know-how
with other group members.
3. If it were entirely up to me, I am confident that I would be able
to share my working experience or know-how with other group
members.
External Control (Armitage et al., 1999)
1. Whether I share my working experience or know-how with other
group members is entirely up to me.
2. I feel there is no obstacle to share my working experience or
know-how with other group members.
3. I feel I often have no opportunity to share my working experience
or know-how with other group members.

Appendix B. Covariance matrix

ABT1 ABT2 ABT3 ABT4 ABT5 SHV1 SHV2 SHV3 SHV4 TKSI1 TKSI2 TKSI3 TKSB1 TKSB2 TKSB3 ITC1 ITC2 ITC3

ABT1 3.928
ABT2 3.207 4.487
ABT3 2.763 3.482 6.321
ABT4 2.917 3.732 4.120 4.900
ABT5 2.986 3.735 4.350 3.931 5.431
SHV1 2.078 2.262 2.933 2.536 2.916 4.409
SHV2 2.121 2.219 2.445 2.478 2.539 3.189 4.179
SHV3 2.218 2.321 2.547 2.493 2.429 3.069 3.405 4.216
SHV4 2.076 2.099 2.550 2.298 2.641 2.870 2.210 2.170 4.399
TKSI1 0.206 0.324 0.159 0.260 0.150 −0.089 0.089 0.115 0.072 0.918
TKSI2 0.214 0.273 0.181 0.271 0.167 −0.134 0.049 0.079 0.003 0.755 0.864
TKSI3 0.235 0.293 0.244 0.295 0.154 −0.131 0.067 0.193 0.034 0.757 0.688 0.915
TKSB1 1.939 1.994 1.892 2.099 1.830 1.716 2.095 2.242 1.671 0.167 0.191 0.158 4.708
TKSB2 1.928 1.956 1.718 2.127 1.867 1.592 1.957 2.266 1.629 0.202 0.243 0.224 4.032 4.392
TKSB3 1.894 1.911 1.889 2.121 1.737 1.841 2.155 2.335 1.835 0.043 0.110 0.093 3.984 3.886 4.511
ITC1 0.194 0.464 0.581 0.643 0.573 0.213 0.320 0.268 0.165 0.344 0.320 0.386 0.111 0.097 0.142 1.508
ITC2 0.152 0.393 0.558 0.550 0.486 0.189 0.259 0.229 0.140 0.402 0.386 0.442 0.120 0.080 0.099 1.315 1.460
ITC3 0.201 0.492 0.532 0.598 0.610 0.208 0.293 0.259 0.239 0.486 0.459 0.505 0.205 0.220 0.185 1.028 1.089 1.405
218 S.-C. Yang, C.-K. Farn / International Journal of Information Management 29 (2009) 210–218
References
Adler, P. S., & Kwon, S. W. (2002). Social capital: Prospects for a new concept. Academy
of Management Review, 27(1), 17–40.
Ajzen, I. (1985). From intention to behavior: A theory of planned behavior. In J. Kunl
& J. Beckman (Eds.), Action control: From cognitions to behaviors (pp. 11–39). New
York: Springer.
Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human
Decision Processes, 50, 179–211.
Ajzen, I. (2002). Perceived behavior control, self-efficacy, locus of control, and the
theory of planned behavior. Journal of Applied Social Psychology, 32, 665–683.
Alavi, M. (2000). Managing organizational knowledge. In R. W. Zmud (Ed.), Fram-
ing the domains of IT management: Projecting the future . . . through the past (pp.
15–28). Cincinnati, OH: Pinnaflex Educational Resources.
Alavi, M., & Leidner, D. E. (2001). Knowledge management and knowledge manage-
ment systems: Conceptual foundations and research issues. MIS Quarterly, 25(1),
107–136.
Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice:
A review and recommended two-step approach. Psychological Bulletin, 103(3),
411–423.
Armitage, C. J., & Conner, M. (2001). Efficacy of the theory of planned behavior: A
meta-analytic review. British Journal of Social Psychology, 40, 471–499.
Armitage, C. J., Conner, M., Loach, J., & Willetts, D. (1999). Different perceptions of
control: Applying an extended theory of planned behavior to legal and illegal
drug use. Basic and Applied Social Psychology, 21(4), 301–316.
Becerra-Fernandez, I., & Sabherwal, R. (2001). Organizational knowledge manage-
ment: A contingency perspective. Journal of Management Information Systems,
18(1), 23–55.
Berman, S. L., Down, J., & Hill, C. W. L. (2002). Tacit knowledge as a source of compet-
itive advantage in the National Basketball Association. Academy of Management
Journal, 45(1), 13–31.
Bock, G. W., Zmud, R. W., Kim, Y. G., & Lee, J. N. (2005). Behavioral intention for-
mation in knowledge sharing: Examining the roles of extrinsic motivators,
social–psychological forces, and organizational climate. MIS Quarterly, 29(1),
1–26.
Boland, R. J., & Tenkasi, R. V. (1995). Perspective making and perspective taking in
communities of knowing. Organization Science, 6(4), 350–372.
Bolino, M. C., Turnley, W. H., & Bloodgood, J. M. (2002). Citizenship behavior and
the creation of social capital in organizations. Academy of Management Review,
27(4), 505–522.
Brashear, T. G., Boles, J. S., Bellenger, D. N., & Brooks, C. M. (2003). An empirical
test of trust-building processes and outcomes in sales manager–salesperson
relationships. Journal of the Academy of Marketing Science, 31(2), 189–200.
Burt, R. S. (1997). The contingent value of social capital. Administrative Science Quar-
terly, 42, 339–365.
Carmines, E. G., & McIver, J. P. (1981). Analyzing models with unobserved variables:
Analysis of covariance structures. In G. W. Bohrnstedt & E. F. Borgatta (Eds.),
Social measurement: Current issues (pp. 65–115). Newburry Park, CA: Sage.
Choi, B., & Lee, H. (2003). An empirical investigation of KM styles and their effect on
corporate performance. Information & Management, 40, 403–417.
Chou, T. C., Chen, J. R., & Pan, S. L. (2006). The impacts of social capital on information
technology outsourcing decisions: A case study of a Taiwanese high-tech firm.
International Journal of Information Management, 26, 249–256.
Cohen, P., Cohen, J., Teresi, J., Marchi, M., & Velez, C. N. (1990). Problems in the
measurement of latent variables in structural equations causal models. Applied
Psychological Measurement, 14, 183–196.
Conner, M., & Armitage, C. J. (1998). Extending the theory of planned behavior: A
review for further research. Journal of Applied Social Psychology, 28, 1429–1464.
Conner, M., & McMillan, B. (1999). Interaction effects in the theory of planned behav-
ior: Studying cannabis use. British Journal of Social Psychology, 38, 195–222.
Fahey, L., & Prusak, L. (1998). The eleven deadliest sins of knowledge management.
California Management Review, 40(3), 265–276.
Fitch, J. L., & Ravlin, E. C. (2005). Willpower and perceived behavioral control: Influ-
ences on the intention–behavior relationship and postbehavior attributions.
Social Behavior and Personality, 33(2), 105–124.
Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unob-
servable variables and measurement error. Journal of Marketing Research, 28,
39–50.
Granovetter, M. S. (1995). Getting a job: A study of contacts and careers (2nd ed.).
Chicago: University of Chicago Press.
Hogel, M., Weinkauf, K., & Gemuenden, H. G. (2004). Interteam coordination, project
commitment, and teamwork in multiteam R&D projects: A longitudinal study.
Organization Science, 15(1), 38–55.
Hsu, I. C. (2006). Enhancing employee tendencies to share knowledge—Case studies
of nine companies in Taiwan. International Journal of Information Management,
26, 326–338.
Jöreskog, K. G., & Sörbom, D. (1989). LISREL 7 user’s reference guide. Chicago, IL:
Scientific Software.
Kaplan, D. (2000). Structural equation modeling: Foundations and extension. Thousand
Oaks, CA: Sage.
Käser, P. A. W., & Miles, R. E. (2002). Understanding knowledge activists’ successes
and failures. Long Range Planning, 35, 9–28.
Kline, R. B. (2005). Principles and practice of structural equation modeling (2nd ed.).
New York: The Guilford Press.
Koskinen, K. U., Pihlanto, P., & Vanharanta, H. (2003). Tacit knowledge acquisition and
sharing in a project work context. International Journal of Project Management,
21, 281–290.
Kuhl, J., & Beckmann, J. (Eds.). (1985). Action control: From cognition to behavior. Berlin,
Germany: Springer-Verlag.
Leana, C. R., & Van Buren, H. J., III. (1999). Organizational social capital and employ-
ment practices. Academy of Management Review, 24, 538–555.
Lewis, J. D., & Weigert, A. (1985). Trust as a social reality. Social Forces, 63(4),
967–985.
Liao, S., Shao, Y. P., Wang, H., & Chen, A. (1999). The adoption of virtual bank-
ing: an empirical study. International Journal of Information Management, 19,
63–74.
McAllister, D. J. (1995). Affect- and cognition-based trust as foundations for inter-
personal cooperation in organizations. Academy of Management Journal, 38(1),
24–59.
McEvily, B., Perrone, V., & Zaheer, A. (2003). Introduction to the special issue on trust
in an organizational context. Organization Science, 14(1), 1–4.
Morgan, R. M., & Hunt, S. D. (1994). The commitment–trust theory of relationship
marketing. Journal of Marketing, 58, 20–38.
Nahapiet, J., & Ghoshal, S. (1998). Social capital, intellectual capital, and the organi-
zational advantage. Academy of Management Review, 23(2), 242–266.
Nonaka, I. (1994). A dynamic theory of organizational knowledge creation. Organi-
zation Science, 5(1), 14–37.
Osterloh, M., & Frey, B. S. (2000). Motivation, knowledge transfer, and organizational
forms. Organization Science, 11(5), 538–550.
Polanyi, M. (1967). The tacit dimension. London: Routledge and Keoan Paul.
Putnam, R. D. (1995). Bowling alone: America’s declining social capital. Journal of
Democracy, 6, 65–78.
Segars, A. (1997). Assessing the unidimensionality of measurement: A paradigm and
illustration within the context of information systems research. Omega, 25(1),
107–121.
Sharma, S. (1996). Applied multivariate techniques. New York: John Wiley & Sons.
Simons, T., Pelled, L. H., & Smith, K. A. (1999). Making use of difference: Diversity,
debate, and decision comprehensiveness in top management teams. Academy of
Management Journal, 42(6), 662–673.
Szymanski, D. M., Troy, L. C., & Bharadwaj, S. G. (1995). Order of entry and business
performance: An empirical synthesis and reexamination. Journal of Marketing,
59, 17–33.
Terry, D. J., & O’Leary, J. E. (1995). The theory of planned behavior: The effects of
perceived behavioral control and self-efficacy. British Journal of Social Psychology,
34, 199–220.
Triandis, H. C. (1977). Interpersonal behavior. Psychology & Health, 15, 383–393.
Tsai, W., & Ghoshal, S. (1998). Social capital and value creation: The role of intrafirm
networks. Academy of Management Journal, 41(4), 464–476.
Shu-Chen Yang is an assistant professor in the Department of Information Man-
agement at the National University of Kaohsiung in Taiwan, ROC. His teaching and
research interests focus on knowledge management, consumer behavior and e-
business. His major publications have been in the areas of management of IS and
consumer behavior.
Cheng-Kiang Farn is a professor in the Department of Information Management at
the National Central University in Taiwan, ROC. His teaching and research inter-
ests focus on electronic business, SCM and knowledge management. His major
publications have been in the areas of electronic business and management of IS.
Cheng-Kiang is also a consultant to many government agencies and enterprises.