Figure 2: Summary of relationship between transactive memory ...

jaspersugarlandΛογισμικό & κατασκευή λογ/κού

14 Δεκ 2013 (πριν από 3 χρόνια και 10 μήνες)

94 εμφανίσεις

1


An integrative model of
the role of
trust in transactive memory development



by



Dr Melanie Ashleigh





Dr Jane Prichard

School of Management





School of Social Sciences

University of Southampton




University of Southampton

Highfield, Southampto
n





Highfield, Southampton

SO17 1BJ






SO17 1BJ



mja@soton.ac.uk






J.S.Prichard@soton.ac.uk

+44 (0)23 80594738






+44 (0)23 80597242



Abst
ract


This paper extends Transactive Memory (TM) theory

as it is currently conceptualiz
ed. We propose a
new integrative model of the relationship between Transactive Memory System

(TMS) development and
trust. By utiliz
ing the TM encoding cycle,

the model
proposes that trust acts as an antecedent of TMSs
and

that

wider perceptions of team members’ trust
worthiness

(benevolence and integrity)

also
affect the
development and maintenance of effective TM
Ss in teams. Our conceptualiz
ation considers
the effect of
trust

on both the knowledge

structu
re

and
the transactive processes involved in TMSs. From our
analysis we provide a number of

propositions

and hypotheses relating to different stages of TMS
development to be pursued by fu
t
ure

research.
Finally, we conside
r the managerial implications of our
model.


Keywords: Transactive Memory, Trust, Encoding Cycles, Knowledge Sharing





2


Transactive memory systems (TMSs) are specialized systems of distributed knowledge which enable
the cognitive division of labor for lea
rning, remembering and communicating information about a task
(Hollingshead, 2000; Wegner, 1986). As such TMSs form an integral part of an organization’s wider
knowledge management system and are argued to be fundamental to competitive advantage (Nonaka
&
Krogh, 2009). Consistent with this, research has evidenced the positive impact of TMSs on team
performance through their organizing effect on the knowledge held within teams (e.g. Austin, 2003;
Lewis, 2003; Liang, Moreland & Argote, 1995; Moreland & Myasko
vsky, 2000; Peltokorpi, 2008;
Prichard & Ashleigh, 2007) by limiting cognitive load (Hollingshead, 1998; Prichard, Bizo & Stratford,
2011) and increasing the effective use of unique knowledge in decision making (Brodbeck, Kerschreiter,
Mojzisch & Schulz
-
Ha
rdt, 2007; Stasser, Stewart & Wittenbaum, 1995). Although the role of TMSs in
team performance is well established, currently there is a lack of understanding surrounding both the
antecedents and the factors that influence their on
-
going development
. This

lack of understanding will
inevitably limit managers’ ability to purposively promote TMSs and so team performance.

In this paper
we address this gap in the literature by considering the role of trust.

Trust is central to the way in which TM has been
oper
ationalized



teams with an intact TMS will be
more trusting of others’ task
-
related ability (Liang et al., 1995; Moreland & Myaskovsky, 2000). From this
position trust has only been considered as providing evidence that a TMS is operating effectively, and

then only in terms of
trust developed from beliefs about the trustworthiness of other’s task
-
related ability.
In this paper, however, we argue that the role of trust in TMSs is more extensive

than currently
acknowledged by TMS research
,
and is crucial

to
developing

these knowledge net
works within and
across teams
(
Prichard & Ashleigh, 2007). Furthermore, whilst perceptions of another’s ability will
undoubtedly be important in a TMS, we propose that so too will be other perceptions of trustworthiness

such a
s whether people will behave benevolently and with integrity

(Mayer et al., 1995). Consequently,
we argue
that trustworthiness and trust play

vital role
s

in TMS development and
their

influence needs to
be more fully understood if the rewards of using such
systems in work teams are to be realized.

3


In this paper we advance the TM literature by explicating the role of
trust and trustworthiness in

TMS
development. Specifically, we break away from the current narrow focus on trust considered
only

as
a
reflection

of
a functioning TMS and
only

in terms of
perceptions of

others’ abilities. We clarify
understanding of the relationship between TMSs and trust by presenting an integrative model that aims
to show that (1) trust is an antecedent of TMS development and (2)

the
wider perceptions of
trustworthiness which act as antecedents of trust must

be considered in structuring and using TMSs. In
the rest of this article we first introduce the key constituents of TMS and the sequence through which it
develops. Second, we
define and scope out the conceptualizations of
trust and trustworthiness useful

for considering
their relationship with

TMSs. Finally, we draw on insights from the trust literature and
bring this into the TMS literature to model their interrelationship by
providing a set of propositions and
examples of specific hypotheses that could be pursued by future research.


Transactive Memory Systems

Transactive Memory has been defined in the literature as the combination of the knowledge held by
individual members o
f a group, plus an awareness of who knows what (Wegner, 1986; 1995). It is a
concept that emphasizes the task
-
related expertise in a team and thus can be considered as a subset of
the broader concept of a team mental model (
Mohammed & Dum
ville, 2001). Howe
ver, unlike team
mental models which tend to focus on the ‘sharedness’ of knowledge, in other words what members
hold in common, TM focuses on distributed knowledge.

The literature makes a distinction between TM being the memory that is held within the g
roup, and a
TMS, which describes how members actively use this TM to co
-
operatively encode, store and retrieve
information about complex interdependent tasks (Brandon & Hollingshead, 2004; Lewis, 2003; Staples
& Webster, 2008). TMSs enable a team to alloc
ate information between members and, through
knowing which individual member has expertise about a particular issue, facilitate access to that
information during task performance (Faraj & Sproull, 2000). The development and use of a TMS
4


therefore requires
both the establishment of a network of expertise (including a shared understanding of
where expertise is located) and the deployment of various coordination or ‘transactive’ processes
(Wegner, Giuliano, & Hertel, 1985) that facilitate access to that expert
ise for the purposes of the task.

A TMS is operationalized around three
sets of behaviours

which reflect its structuring and
coordinating constituents; memory differentiation, task coordination and task credibility (Liang et al.,
1995; Moreland & Myaskov
sky,
2000). These have been used as behavioral measures which one would
expect to be present where TMSs are operating to encode, store and retrieve information to perform the
task ( Moreland & Myaskovsy, 2000). Memory

differentiation is the degree to which

individual members
of the team specialize in remembering different aspects of the task. Task coordination is team members’
ability to work together effectively whilst carrying out the task. Finally, task credibility is the degree to
which one team member
perceives that another member has the task expertise or ability to accomplish
their role in the task. The term ‘task credibility’ is synonymous with the concept of trust in ability (Mayer,
Davis & Schoorman,
1995). Whilst memory differentiation has been ma
pped clearly to the structuring
constituent of TMS and task coordination to the transactive process constituent, the role of task
credibility has been less clearly explicated by the literature, appearing to be a behavior that evolves
from TMS developing wi
thin the group. It

is a competence
-
based trust resulting from confidence
between members who know the location of expertise within the team. When team members are aware
of each other’s task knowledge they will be more likely to accept others’ contributions
, be less critical of
others and be less likely to make claims of expertise (Austin, 2003; Liang et al., 1995; Rulke & Rau,
2000). This reliance is evidence of trust developing within the relationship as a consequence of
knowledge about expertise being sha
red between team members when utilizing the TMS. Theorizing
about the relationship of TM and trust in work teams has not extended much beyond this point.

Based on their observations of teams performing an interdependent task, Rulke and Rau (2000)
propose t
hat TM develops through a series of encoding cycles during which members seek to establish
the location of expertise within the group (see central circle of Figure 1). Each encoding cycle begins
5


with members requesting information about a specific area of
the task. In response, knowledgeable
members make claims of expertise in that area whilst unknowledgeable members declare that they do
not hold the necessary expertise. These claims are subsequently evaluated by other members in order
to identify the knowl
edge distribution within the group. The next stage of the cycle involves the
coordination of task responsibilities based around the distribution of expertise identified. Finally, task
performance feeds into future evaluations of member expertise and the cy
cle begins again. Thus each
cycle is a sequence of communication and interaction between members, described by Wegner,
Giuliano and Hertel (1985) as transactive processes, which enable the encoding, storing and retrieval of
knowledge.

Rulke and Rau (2000)

found that interactions early in the task are marked by members claiming
those task areas where they have specific expertise; however this usually leaves some task domains
unclaimed. Subsequent encoding cycles therefore establish where gaps lie in the kno
wledge base, from
which members then allocate responsibility to those regarded as best able to fill them. Rulke and Rau’s
findings are further supported by Lewis (2004) who found that the planning stage in team activity was
critical to the development of T
MSs. She argued that frequent interaction during this stage helped to
develop accurate perceptions of knowledge distribution and how members’ different areas of expertise
fitted
together. Essentially, the different stages of the TM encoding cycles combine
to form an
information
-
processing framework that models the development of a TMS; a distributed knowledge
system which describes how information is encoded, stored and retrieved over time within teams.


Trust

Mayer et al. (1995) have defined trust as “a w
illingness to be vulnerable to the actions of another
party based on the expectation that the other will perform a particular action important to the trustor,
irrespective of the ability to monitor or control that other party” (p. 712). The decision to tru
st (or not) is
based on the trustor’s perceptions of the characteristics of the trustee together with their own propensity
6


to trust (Mayer et al., 1995), and is argued to have cognitive, affective and behavioral components (e.g.
Cummings & Bromiley, 1996;
Lewis & Weigert, 1985). Evaluating perceptions of others’ trustworthiness
and the level of risk required in considering how to behave is a rational, cognitive decision. However,
emotion also influences the decision to trust as
feelings, both positive and n
egative, are associated with
how perceptions of trustworthiness are formed, thus influencing behavioral intentions
(Schoorman,
Mayer & Davis, 2007)
.
Furthermore, the outcomes of trusting another will influence the way in which
people think and feel about t
rusting in the future in a continuous dynamic cycle.

Perceptions of trustworthiness are antecedents of trust and are themselves based on beliefs across
three distinct but related factors (Mayer et al., 1995); the trustee’s ability (their domain
-
specific
kn
owledge), their benevolence (the extent to which the trustor perceives the other will act in their best
interests), and their integrity (whether they are principled and likely to keep their promises). Such beliefs
may be based on previous experience or alt
ernatively on various trusting bases such as social
comparisons and stereotypes, associated with cognitive cues such as race, age and professional
identity (Hogg & Terry, 2000; Li, Hess & Valacich, 2008; McKnight, Cummings & Chervany, 1998;
Meyerson, Weick

& Kramer, 1996). These in turn trigger expectations about others’ attributes such as
status, role, ability, honesty and cooperativeness (Chatman & Flynn, 2001; Levin, Whitener & Cross,
2006; McAllister, 1995). The role of all these factors of trustworthin
ess in developing trust highlights that
focusing only on trust in others’ ability (as is currently the case in TM literature) ignores the important
contribution of beliefs about their benevolence and their integrity. Work teams do not operate in a social
v
acuum in which individuals focus only on others’ expertise, but rather they are continually affected by a
range of social factors that influence beliefs about the motivations of others to engage in knowledge
-
sharing activities.

To
further
inform the relat
ionship between
trust

and

TMS, it is useful to refer to the ideas of McEvily
et al. (2003) who have proposed that the notions of trust and trustworthiness can be seen as organizing
principles of workplace interactions and processes. They argue that
trust b
ased on perceptions of
7


trustworthiness can be viewed as both a structuring device which shap
es the networks of interactions
between people (for example by influencing the degree of role specialization), and as a mobilizing
device that enables and constrain
s the coordination of work (for example, knowledge sharing,
monitoring and safeguarding behaviors). These two organizing effects of trust are particularly useful for
thinking about TMS development as they map directly onto, and so are likely to influence,
the two key
constituents of TMS; the network of knowledge (the structure), and the transactive processes that
enable the coding storage and retrieval of that knowledge (mobilizing processes).

This brief summary of
trust and trustworthiness is

not intended

to provide an exhaustive review of the
trust literature, but rather to offer a basis for thinking about how
these

influence TMS development. The
organizing effects of trust identified by McEvily et al. (2003) suggest that it is likely to be an important
a
ntecedent to TMS development, both in terms of establishing the knowledge network and mobilizing
the transactive processes needed to encode, store and retrieve knowledge from that network to perform
a task. Additionally, Mayer et al.’s (1995) factors of tr
ustworthiness that underlie the decision to trust
show that the TM literature should further include consideration of beliefs about benevolence and
integrity. Indeed, as we will discuss, trust in another’s ability alone may be of little value if team
membe
rs do not also trust in the integrity of others or in their benevolence towards the team and
achievement of group rather than individual goals.


Modelling the TMS
-
Trust Relationship

To address our aims and model the relationship between TM and trust, we st
ructure our discussions
using Rulke and Rau’s (2000) encoding cycle framework to deconstruct the stages of TMS development.
This framework proposes that TMS develops through four stages; making claims of expertise (stage1);
evaluating expertise (stage 2);
coordinating task performance based on the expertise distribution
identified (stage 3), and reviewing and refining expertise distribution to aid future performance (stage 4).
Although in our analysis we articulate the role of trust at each stage of this cy
cle as a discrete and
8


sequential process to aid explication, we recognize that in reality the stages of the TM encoding cycle
are more interdependent and continuously evolving (Brandon & Hollingshead, 2004;
Lewis & Herndon,
2011;
Rulke & Rau, 2000).

In co
nsidering trust as an antecedent of TMS (Aim 1), in our analysis we develop propositions of the
role of trust at each stage of the encoding cycle. In incorporating wider perceptions of trustworthiness
other than in ability (Aim 2), we consider how beliefs
about benevolence and integrity also influence
TMS development at each stage an
d

develop a set of propositions showing the importance of all three
factors of trustworthiness. It is not our intention however to map each separate factor of trustworthiness
on
to each stage of the TMS. Indeed, such an undertaking may be impossible given that each factor of
trustworthiness operates along a continuum from low to high, and may vary in relation to one another
(Mayer et al., 1995). Such properties result in infinite
possibilities as to how these factors might combine
to influence the level of trust and TMS development. Furthermore, in many cases, the effects of
benevolence, integrity and ability may have the same outcome on TMS development, albeit for different
reason
s. Finally, in relation to both aims we present sample hypotheses of how trust may influence the
two constituents of TMS; the structure of the knowledge network and the transactive processes needed
to encode, store and retrieve knowledge. Figure 1 illustra
tes these relationships and summarizes the
propositions.

Insert Figure 1 about here


To achieve the aims of this paper, we consider a team with no history of working together and therefore
no existing TMS. This will enable us to examine the development of
TMS from the very beginning
through to its establishment as a system for managing distributed knowledge in the performance of
ongoing work cycles (Bradley, White & Mennecke, 2003; De Jong & Elfring, 2010). To enable a broad
focus around trust, we specifica
lly exclude from our analysis consideration of temporary teams where
research has proposed that, because members have no expectations of working together in the future,
9


a qualitatively different type of
swift

trust will develop based on a narrower set of t
rust parameters than
found in other teams (Meyerson et al., (
1996). Thus we consider a team which will be ongoing but
otherwise do not set any specific boundaries at this stage; however, in our conclusion we discuss
various conditions that may affect the T
MS
-
trust relationship including the implications of leadership and
variations in team contexts.


Stage 1: Making Claims of Expertise

When the members of a newly formed team initially come together, they will have limited knowledge
about one another. Betwe
en them, however, they will possess various pieces of
lower
-
order information
;
bits of expertise that will be relevant to the task in hand (Wegner et al., 1985). In the first stage of the
encoding cycle members seek to establish what lower
-
order informatio
n is present and where in the
team it is located. In response to requests for such information, if TM is to be developed, team members
must declare their own level of expertise, either claiming knowledge or admitting a lack of it. We
propose that trust in
others is necessary if such information is to be released to them.

The trust and knowledge management literature has previously established that trust is an antecedent
of knowledge sharing (e.g. Dirks & Ferrin, 2001; Mayer et al., 1995; Sharkie, 2005) dem
onstrating that
trust promotes knowledge sharing and consequent group performance (Koskinen, Pihlanto &
Vanharanta, 2003; Mishra, 1996; Nelson & Cooprider, 1996; Politis, 2003; Prichard & Ashleigh, 2007;
Zand, 1972). Conversely, knowledge sharing is likel
y to be restricted when trust is absent (
Madhava
n &
Grover, 1998). Without trust people may hesitate to transfer knowledge within a group for many
different reasons (Kim & Mauborgne, 1998). For example, members may not feel psychologically safe or
comfort
able enough to risk divulging information for fear of making errors and being rejected by other
team members (Edmondson, 2003). They may fear losing face; that the recipient of their knowledge
may abuse it in some way, or that their knowledge contributions

may be misleading, inaccurate or
irrelevant to the discussion (Ardichvili, Page & Wentling, 2003). Curşeu and Schruijer (2010) found that
10


low trust can lead people to feel attacked when releasing information, increasing the probability of
conflict which i
n turn makes it less likely that knowledge release will continue. Thus, trust at this stage
will both increase psychological safety and reduce intra
-
group conflict, positively impacting on the
declaration of knowledge.
The feelings which are inevitably ass
ociated with beliefs such as
psychological safety or conflict mean that trust will not only be influenced by cognitive evaluations of
team members’ trustworthiness, but also by emotional aspects that impact on the decision to trust and
therefore whether to

release or retain information.

With previous literature supporting the notion that trust is needed for knowledge sharing to occur, the
next issue to consider at this stage is whether initial trust can exist between unfamiliar team members to
allow the en
coding cycle to begin. When a new team is formed, there is no direct knowledge or first
-
hand experience of other team members on which to base perceptions of trustworthiness. Consequently,
decisions about whether to engage in trusting behaviors such as sha
ring knowledge need to be made.
Theoretical frameworks in the trust literature have proposed the existence of a number of alternative
‘trusting bases’ that serve as a proxy for experience and can facilitate trust between team members.
Previous research has

proposed that a baseline of moderate to high trust could exist due to personality
(e.g. Cattell, 1965; Gill, Boies, Finegan & McNally, 2005: Mayer et al., 1995), cognitive processes (e.g.
Li et al., 2008; Meyerson et al., 1996), or organizational culture
(e.g. Schoorman et al., 2007; Sitkin,
1995).

These trusting bases evoke both cognitive and affective evaluations of each of the factors of
trustworthiness and will determine (at least in part) whether or not one party will trust another, and
therefore whet
her they will choose to release or withhold knowledge.

This combined literature demonstrates that trust is an antecedent of knowledge sharing and that it is
possible for trust to exist when a team first comes together. Since the disclosure of knowledge is
a
crucial transactive process at stage 1, the existence of trust is therefore essential at this point for
knowledge to be released in order to initiate the development of the knowledge
-
sharing network. In the
absence of trust, people will be reluctant to r
isk divulging knowledge; instead limiting the amount,
11


completeness or accuracy of what is shared, which in turn would prevent the completion of this stage.
Furthermore, trust at this stage is dependent on perceptions of benevolence, integrity and ability
e
stablished from the range of trusting bases that may be operating in any given context.
Although
perceptions of trustworthiness of others’ ability and benevolence may feature at this point, Mayer at al.
(1995) suggest that perceptions of another’s integrit
y will be especially important at this early stage. If
the knowledge receiver is perceived as being open and honest then the claimant will be more likely to
take the risk of releasing information. This type of openness is essential for mobilizing team
perf
ormance because it increases the motivation for knowledge sharers to reveal unique knowledge
which can subsequently be recombined in novel ways to benefit the task in hand (Mayer et al., 1995;
McEvily et al., 2003).

Many propositions can be derived from t
he previous discussions about the effect of trust on the
disclosure of knowledge at stage 1 of the TM encoding cycle. However, two general assumptions that
can be made in relation to our two aims are; that trust must be present for the full disclosure of
k
nowledge at this stage and that perceptions of others’ trustworthiness which underlie the decision to
trust will not simply relate to their ability but additionally to their integrity and benevolence.


Proposition 1a: Trust is an antecedent which promotes

the declaration of knowledge by team
members at stage 1 of the TM encoding cycle.

Proposition 1b: Declarations of
knowledge by team members at stage 1 of the TM encoding cycle
will be influenced by trust based on perce
ptions of the three factors of trus
tworthiness: benevolence,
integrity and ability.

From these propositions it is possible to develop a number of hypotheses to consider how trust might
influence the two constituents of a TMS; the knowledge structure and the transactive processes. At this
st
age there will be little effect of trust on the structure of the network as the knowledge is yet to be
12


evaluated, but there will be a clear effect on the transactive process of declaring knowledge. Therefore
a sample hypothesis might be:

H1: Negative perce
ptions of benevolence, integrity and/or ability will limit trust, so decreasing the
amount of knowledge declared at stage 1 of the TM encoding cycle.

Negative perceptions of trustworthiness across any of these dimensions will adversely affect how
knowledge

is disclosed at this stage, either in terms of how much information is disclosed, or its quality
in terms of completeness or accuracy. This hypothesis could be tested experimentally by manipulating
people’s perceptions of the presence/absence of others’ t
rustworthiness at this early stage by varying
information provided about an unfamiliar team member’s reputation, or by making salient the differences
in people’s in
-
group memberships.

Stage 2: Evaluating Claims of Expertise

In the second stage of the enco
ding cycle the information disclosed at stage 1 is evaluated to
establish the distribution of existing expertise, to identify where any potential gaps may lie, and to
establish whether claims made can be relied upon as the basis for subsequent role allocat
ion. Wegner
(1995) proposed that in the absence of personal experience, assumptions about whether someone has
expertise can come from four possible sources; surface level characteristics (gender, ethnicity, etc.); the
assignment of a team member to a speci
fic role by a manager; when and for how long someone claims
an area of expertise, and expertise based on occupational roles or qualifications. We argue that
regardless of which of these is used to evaluate expertise, cognitive and affective appraisals of t
he
benevolence, integrity and ability of others will all play an important role in whether these evaluations
are positive or negative. For example, stereotypes about surface characteristics such as age, or
heuristics associated with particular qualificatio
ns, will influence perceptions of ability. Similarly,
assumptions based around a shared group membership may influence perceptions of benevolence and
13


whether others are committed to the
team. These in turn will influence the decision to trust, or not, the
information released by others.

Perceptions about the trustworthiness of others arising from such trusting bases will be further
influenced by the trusting behaviors displayed by others at stage 1. Knowledge shared freely at stage 1
will reinforce the beli
ef that others are trustworthy, thus creating an upward spiral of trust (Ferrin, Bligh &
Kohles, 2008; Zand, 1972). Alternatively, if knowledge is withheld by others at stage 1, this may
disconfirm positive beliefs about their trustworthiness and so limit
trust. In summary therefore, each of
the three factors of trustworthiness developed from the various trusting bases and revised by prior
experience at stage 1 will influence beliefs about whether a team member can be trusted to take
responsibility for a sp
ecific area of expertise.

Ardichvili et al. (2003) support that trust is an important precursor for team members to believe that
the source of knowledge is reliable and has provided objective information. Furthermore, a high level of
trust between people i
s also likely to speed up this stage of the encoding cycle (McEvily et al., 2003). In
considering the organizing effects of trust on knowledge sharing more generally, McEvily et al. (2003)
argued that high trust mobilizes the coordination of resources by m
aking it more likely that people will
take that information at ‘face value’. In other words, trust acts as a heuristic in the decision
-
making
process, making it less likely that evaluations will be made through prolonged periods of questioning
and discussi
on aimed at establishing reliability. Furthermore, high trust may increase the likelihood that,
even if people are uncertain about the claims made, they will suspend their judgment of others, instead
giving them the benefit of the doubt rather than engagin
g in safeguarding behaviors (McEvily et al.,
2003). Where safeguarding activities are low the transaction costs associated with getting the work
done will be lower; for example, less time spent evaluating the claims of a trusted team member frees
time to e
valuate the claims of others. Thus, trust influences a range of transactive processes necessary
for the evaluation of knowledge at this stage. Of course there are risks related to trust being too high at
14


this stage as it may reduce the evaluation process o
f claims made by others which are in fact incorrect
or misleading (Langfred, 2004).

Trust in the claims of expertise made by a team member will also inevitably alter the expertise
-

location information held by the team with implications for subsequent sta
ges of the encoding cycle. For
example, a lack of trust would make one question the validity or accuracy of claims of expertise and so
(rightly or wrongly) structure beliefs about the location of information in ways that would limit a distrusted
member’s c
entrality in that domain. Discounting claims of expertise may potentially lead to gaps in the
knowledge network which may be problematic for performance if these gaps are not identified and
rectified. So in this way, trust structures the network in terms o
f role specialization across the team
(McEvily et al., 2003).

Thus the presence or absence of trust based on the three factors of trustworthiness will not prevent
the evaluation of knowledge taking place as this will happen regardless; however, it
will

af
fect whether
that knowledge is perceived in a positive or negative way.
Hence the greater the trust between team
members the more positively claims of expertise will be evaluated.
Relating this to the two aims of this
paper, our analysis leads us to our se
cond set of propositions:


Proposition 2a: Trust is an antecedent which promotes positive evaluations by team members of
claims of expertise at stage 2 of the TM encoding cycle.

Proposition 2b: Evaluations by team members of claims of expertise made at st
age 1 will be
influenced by trust based on perceptions of the three factors of trustworthiness: benevolence,
integrity and ability.

From these propositions it is possible to develop a number of hypotheses relating to how trust might
influence the two const
ituents of a TMS. As we have discussed in our analysis, at this stage trust will
affect both the emerging structure of the network and also the types of transactive processes and how
15


they are utilized. Two sample hypotheses, one relating to the TMS structu
re and the other relating to
transactive processes, might be:

H2: Perceptions of benevolence, integrity and ability will structure beliefs about the location of
expertise.

H3: Negative perceptions of the benevolence, integrity and/or ability will increase
the time spent
evaluating claims of expertise.

The role that trust plays in structuring beliefs about the location of expertise at this stage might be
demonstrated operationally by comparing individual team members’ mental model of the location of
expertis
e (for example, measured through individual team members’ ratings of others’ expertise) with
the claims that were made at stage 1 of the TM encoding cycle. A high degree of congruence would
indicate a high level of trust. Where trust is low however, claims

of expertise would be less likely to
match individual team members’ models of knowledge distribution.

To summarize, during these first two stages of TMS development, trust acts as an antecedent of
both the structure of the knowledge network and the trans
active processes engaged in. Furthermore,
each of these TMS constituents will be
influenced by trust based on each of the perceptions of
trustworthiness; benevolence, integrity and ability. Thus
, we propose that trust affects team members’
understanding of

the location of expertise in the team; for example, those members whose claims are
trusted will be more central to the network when the expertise they possess is required for the task.
Additionally, trust affects the type of transactive processes deployed

and how these are utilized; for
example, through its influence on suspending judgments about others when either disclosing or
evaluating claims of expertise. It is important to note that location information will not necessarily be
accurate at this stage
as perceptions of trustworthiness are based on proxies such as stereotypes and
reputations and some limited behavior of declaring knowledge, rather than actual performance of the
task.

16



Stage 3: Coordinate Task Based on Expertise Identified

Stages 1 and 2

result in the formation of tentative hypotheses about whether other team members
can take responsibility for specific domains of expertise (Brandon & Hollingshead, 2004). In the third
stage of the encoding cycle, team members use the knowledge distributio
n identified and evaluated in
previous stages to allocate task responsibilities and coordinate activities to perform the task. Crucially,
at this stage, members must be able to retrieve the necessary knowledge from those deemed as
experts and allocate new
information to them as it becomes available. To do this efficiently they must be
able to take the risk of relying on those experts having the ability to perform their allocated roles, feel
confident that their actions will be for the good of the team and d
irected towards the task, and feel
confident that they will use their expertise with integrity, acting fairly and in ways that are consistent with
team expectations. Therefore, as in the previous two stages, all three factors of trustworthiness
influence t
he decision to trust and so

the ongoing development of TMSs.

For example, in the absence of positive perceptions about the ability, benevolence or integrity of
others, team members are likely to spend time excessively checking or monitoring the claimant’s

activities in order to reduce their own vulnerability (McEvily et al., 2003). Although some level of
monitoring or heedful relating is important during task performance to support team processes
(Spreitzer, Sutcliffe, Dutton, Sonenshein, & Grant, 2005), w
hen people overly monitor they spend
relatively less time attending to their own area of responsibility (Cummings & Bromiley, 1996). Thus a
lack of trust will interfere with the efficiencies that would otherwise be afforded by a TMS so preventing
team memb
ers from coordinating the task efficiently. Counteracting such a view are the findings of Rau
(2005) who, despite hypothesizing that trust should enhance team members’ use of location information
about expertise, was unable to demonstrate any association i
n a field study of management teams.
However, the absence of any association here may have been due to the measure of trust which only
measured beliefs about integrity and had been developed to measure employees’ trust in their
17


employers rather than trust
between team members. In contrast, Akgün
,
Byrne, Keskin,

Lynn

&
Imamoglu

(2005) reported that trust did predict task co
-
ordination, indicating empirical support for trust
promoting TM usage.

In addition to being mobilized by trust, it is further likely t
hat interactions at this stage will also lead to
the emergence of trust in a continuous upward spiral between trust and cooperative behavior (Ferrin et
al., 2008; Zand,1972). From an analysis of virtual teams, Curşeu (2006) proposed that early positive
tea
m interactions, which might include role allocation and planning, will lead to the development of trust
which in turn will positively impact on future coordination processes. We return to this issue of the
emergence of trust more fully when considering sta
ge 4.

We note that there are also potential dangers if trust is too high at this stage in that it may stop team
members from monitoring each other’s needs and expectations (Enzle & Anderson, 1993; Langfred,
2004). Langfred (2004) found that high intra
-
team

trust combined with high levels of member autonomy
led to process losses and co
-
ordination errors, negatively affecting performance. This research
demonstrates that a high level of trust is sometimes a disadvantage for teams and that in some contexts
trus
t will have a non
-
linear effect on TMS development. Too little trust in a TMS can result in excessive
time and effort devoted to monitoring the activities of others, whilst too much trust can result in an
absence of heedful relating needed to reduce the po
ssibilities of process losses.

In considering the influence of trust on TMS development it is also important to recognize the
potential for one team member to hold contrasting beliefs about another in terms of their ability,
benevolence and integrity. Alt
hough these three factors of trustworthiness inter
-
relate they are also
separable (Mayer et al., 1995). Thus whilst one team member may have a high level of trust in the
ability of another and a good understanding of how knowledge is distributed between th
em, they may
not rely on the TMS because of concerns about that other’s trustworthiness in terms of their
benevolence or integrity. Under such circumstances, the TMS may be disrupted if the team member
consequently engages in safeguarding behaviors, such a
s developing higher reliance on knowledge
18


held in common, in order to minimize the risk both to themselves and to the task. Such behavior will
increase the level of redundancy in the network and reduce the degree of role specialization.

This analysis in wh
ich positive beliefs exist about the others’ trustworthiness in terms of ability, but
where there is an absence of positive beliefs about their benevolence, challenges a premise currently
assumed across TM research that all team members work towards achiev
ing the same goal. Instead, it
suggests that this assumption may not always be true as it is possible that individuals may have
additional or alternative self
-
motivated goals. We argue that multiple motives are often present in teams
and will impact on the

complexity of the relationship between TMSs and trust (Brandon & Hollingshead,
2004; Peltokorpi, 2008; Rau, 2005). This issue of mixed motives has been the focus of research by
Jarvenpaa and Majchrzak (2008) on knowledge collaboration among professionals
from separate
organizations protecting national security. They proposed that in such contexts experts are not always
free to share all aspects of their knowledge. As a consequence, whilst trust in ability will exist,
benevolent
-
based trust cannot be assume
d. Jarvenpaa and Majchrzak argue that, in such
circumstances, moderate levels of benevolent
-
based
distrust

may be helpful to TMSs as this will result
in confident negative expectations about others’ motives. Such expectations will prevent group
members fro
m relying too heavily on experts by triggering a more mindful evaluation of the information
received from others and the recognition that there is some knowledge that others will not share. The
certainty provided in this regard helps members combine their
own knowledge with that of others in the
network.

Despite the claim that distrust is likely to be a necessary part of mixed motive contexts where it acts
as a protective mechanism against deliberately misleading knowledge transfer, we argue that it is als
o
likely to have a negative impact on all stages of the encoding cycle. Although the TM structure will be
present in terms of the existence of expertise and knowledge of where that expertise lies, distrust will
limit transactive processes and therefore und
ermine the value of a TMS as knowledge shared cannot be
adequately evaluated and retrieved. Consequently, effort will be expended in monitoring and
19


corroborating information received. Over time, such behaviors will increase the amount of knowledge
held in
common, leading team members to become less reliant on the TMS. Although we believe that
distrust will be damaging to TMSs, we agree with Jarvenpaa and Majchrzak (2008) that users of
distributed knowledge systems must be aware of the complex influence of t
he three different elements
of trustworthiness to understand how knowledge will be used in such systems; or indeed whether such
systems will be relied on at all.

In summary, we propose that in stage 3 of the encoding cycle, trust based on the different fac
tors of
trustworthiness (which may themselves vary independently) promotes certain transactive processes (e.g.
safeguarding, reliance) that facilitate or impede access to and use of knowledge about expertise.
Furthermore, this will impact on the structure
of the TMS such that more trusted members develop a
higher degree of specialization, whilst there will be reduced reliance on and more knowledge held in
common with less trusted members. This leads us to our third set of propositions.

Proposition 3a: Trust

is an antecedent which promotes the efficient coordination of responsibilities
between team members based on the expertise distribution identified.

Proposition 3b: The allocation of roles and types of transactive processes deployed by team
members will be

influenced by trust based on perceptions of the three factors of trustworthiness:
benevolence, integrity and ability.


From these propositions and our analysis it is possible to develop a number of hypotheses relating to
how trust might influence the two
constituents of a TMS. Two sample hypotheses, one relating to the
TMS structure and the other relating to transactive processes, might be:

H4. Negative perceptions of benevolence, integrity and/or ability will decrease the congruence
between the actual loc
ation of expertise and how roles are allocated.

20


H5. Positive perceptions of benevolence, integrity and/or ability will decrease monitoring and
safeguarding behaviors.

In order to test such hypotheses, levels of trustworthiness could be manipulated and then

comparisons made between the allocation of roles and actual expertise measured by individual tests of
task knowledge. Similarly, measures of monitoring and safeguarding behaviors could be correlated with
levels of trustworthiness.


Stage 4: Reviewing an
d Refining Expertise Distribution

In the final stage of the encoding cycle, the level and quality of performance on the task to date will
be used to update knowledge about the location of expertise within the team. The refined TMS will then
form the basis
to determine how expertise will be used as the task progresses, or when a new task
begins, and what members will choose to learn about the task in the future. Again trust is important, but
now as well as
promoting
the development of the TMS, in a reciproca
l relationship, it develops
from

the
TMS. If team members successfully execute their responsibilities in their allocated area of expertise, this
will increase the level of trust in each other’s ability. As the task continues and further encoding cycles
are

completed, trust in ability will continue to increase, resulting in the high levels of task credibility
observed in both experimental and field studies of TMSs (Austin, 2003; Lewis, 2004; Liang et al., 1995;
Rulke & Rau, 2000). We propose that this comes
about because during task performance the task
-
oriented interactions, typical of the knowledge sharing and evaluation stages (themselves made
possible by trust), in turn promote trust in others’ abilities in an upward spiral (Curşeu, 2006; Zand,
1972). In
a study of virtual teams, Kanawattanachai and Yoo (2007) found that early and frequent task
-
oriented communication promoted trust in ability. Alternatively, poor task performance may lead to a
withdrawal of trust from some members with subsequent monitorin
g and safeguarding behavior leading
to a downward spiral of trust across successive encoding cycles.

21


Other research suggests that wider perceptions of trustworthiness beyond ability, such as
benevolence and integrity, co
-
evolve with cooperation (e.g. Ferri
n et al., 2008; Mayer et al., 1995). With
cooperation being a necessary aspect

of task coordination, which
itself

is

a central aspect of TMS
operation, these findings support the notion that trust and TMSs develop in co
-
evolutionary ways. Thus
it appears t
hat initial trust extended in the early stages of TMS development, based not around
experience but around various trusting bases, enables knowledge sharing and coordinated activity to
take place, which in turn leads to the development of trust now rooted i
n early task
-
related experiences.

By this stage therefore, in addition to trust influencing TMS development, it is now itself influenced by
the
TMS. This reciprocal relationship between trust and knowledge sharing is supported

both
theoretically and empir
ically by the wider trust literature (Dirks, 2000; Ferrin, Bligh, & Kohles, 2008;
Mayer, Davis, & Schoorman, 1995; Zand, 1972). We argue that this

outcome should not only be
measured in terms of trust in ability, as has previously been the case, but should

also consider
perceptions of benevolence and integrity. Some studies have begun to include these broader
conceptualizations in relation to understanding TMSs (Jarvenpaa & Majchrzak, 2008; Prichard &
Ashleigh, 2007). For example, a study by Prichard and As
hleigh (2007) found a positive relationship
between TM and trust in teams using a multi
-
dimensional measure of trust taken at the end of the task.
The study also measured behavioral variables during task performance including quality of interaction,
and fr
equency and equality of participation, and found that all were positively associated with both
cognitive and affective sub
-
factors of trust where the latter may be considered to map quite closely to
the factors of benevolence and integrity. These results m
ay indicate that the behaviors engaged in
during earlier stages of the encoding cycle lead to the development of trust that extends beyond
cognitive evaluations about others’ abilities.

The review and refinement stage at the end of the cycle serves as a f
eedback loop into successive
encoding cycles as members decide whether to revise the knowledge network. These revisions will be
influenced by the perceptions of trust
worthiness

that have emerged from early TMS development, and
22


will in turn influence both t
he structure and transactive processes in subsequent cycles. If a member’s
perceptions of another’s ability, benevolence and integrity remain high or low at this stage (compared to
early stages), it is likely that

each will retain their own responsibilities in their initial areas of expertise. In
contrast, changes in the level of trust as a result of this stage will lead to a re
-
allocat
ion of roles and so
alter the structure of the knowledge network in the future until a point of stability is reached.

In the longer term, as teams continue their ongoing work cycles, if trust is high then new information
will be directed to the appropriate

expert, so strengthening their knowledge, thus increasing the degree
of specialization within the team. Lewis (2003) however argues that members will only commit to a
distributed system of knowledge sharing if they can rely on other members to remember th
eir own
specific areas of the task. Lewis only considered trust in ability

(e.g. task credibility); however, we argue
that without additional positive beliefs about benevolence and integrity, co
-
workers may aim to
contribute to the team goal in a way that

limits the risk of reliance on others (Dirks & Ferrin, 2001). We
would propose therefore that an absence of positive beliefs about any of the factors of trustworthiness is
likely to impede subsequent TMS development as members will not be prepared to rely

on others taking
responsibility for a given area of expertise in the future. Consequently, members will act to minimize risk
by learning this area of the task themselves with consequential implications for cognitive overload
(Prichard et al., 2011). Thus,

where trust in other team members is low, it is likely that there will be a
higher level of information held “in common” relative to distributed information, so minimizing the
benefits of a TMS.

Therefore, we propose that in stage 4 of the encoding cycle
, revised perceptions of benevolence,
integrity and ability emerging from TMS development will serve as an organizer of both the structure of
the knowledge network and transactive processes involved in future cycles of TMS development. This
leads us to our

fourth set of propositions which reflect trust
developing from

TMS and its role in
successive TMS cycles:


23


Proposition 4a: Effective TMSs promote trust between team members.

Proposition 4b: Effective TMS operation will promote trust between team members
based on
perceptions of the three factors of trustworthiness: benevolence, integrity and ability

Proposition 4c: Trust influences the structure, transactive processes and use of TMSs by team
members in successive encoding cycles

Proposition 4d: The alloca
tion of roles and type of transactive processes deployed in successive
encoding cycles will be influenced by trust based on perceptions of members’ benevolence, integrity
and ability.

From these propositions and our analysis of this stage it is possible to

develop a number of hypotheses
relating to how trust might influence the two constituents of a TMS. Two sample hypotheses, one
relating to the TMS structure and the other relating to transactive processes, might be:

H6: Changes in perceptions of benevolen
ce, integrity and/or ability compared to earlier stages will
result in a redistribution of roles

between team members
.

H7: Stability in perceptions of benevolence, integrity and/or ability will result in stability in transactive
processes
used by team memb
ers
across successive encoding cycles.

The first hypothesis could be measured by comparing the stability of perceptions of trustworthiness
between stages 2 and 4 with the stability of role allocations between stage 3 and the subsequent
encoding cycle. The
second hypothesis could be measured by tracking the stability over time of various
transactive processes such as sharing of vital information between team members and neglect of task
areas due to excessive monitoring, and correlated against the stability o
f measures of trustworthiness.


24


Conclusion

Theoretical contribution

The goal of this article was to extend existing knowledge management literature in relation to TM
theory by incorporating
from the trust literature
a more detailed understanding of the ro
le of trust in TMS
development. Currently its role in TMS development is under
-
emphasized, largely limited to a focus on
cognitive evaluations of trust in ability (i.e. task credibility) found when a TMS is operating effectively. As
such, trust is virtuall
y a ‘silent presence’ in the TM literature which ignores its potential for either
promoting or disrupting TMSs and thus team performance. In response we argue that, in accordance
with our two aims, the role of trust in TM must be extended.

In relation to
our first aim, we propose that trust should be seen as an explicit antecedent in TMS
development. Trust acts to organize TMSs by influencing both the knowledge structure and transactive
processes. In the different stages of the TM encoding cycle, trust act
s to mobilize team members to
contribute information (stage 1); evaluate the information received from others in a positive or negative
light (stage 2); coordinate with others to combine expertise to complete the team task (stage 3), and
review task perfor
mance and redistribute roles where necessary based on expertise ready for the next
cycle (stage 4). Therefore, we argue that trust increases openness in knowledge sharing in the TMS
system allowing it to operate. This ensures that the unique knowledge held

by individual members is
used to promote the quality of team outcomes (Mesmer
-
Magnus & DeChurch, 2009; Stasser et al.,
1995). Furthermore, consistent with the organizing effects of trust (McEvily et al., 2003), we believe that
trust impacts on how knowled
ge is distributed amongst team members by either raising or lowering the
degree of memory specialization. Where trust levels are high and experience of task performance and
working together validates trust in others, we propose that there will be an increa
se in the degree of
specialization within the team. If trust in some members is low, however, it will influence their position in
the TMS such that others will limit their reliance on that person as an expert. Indeed, it may lead to the
team abandoning tha
t member altogether. Our theoretical stance also posits that the level of trust
25


changes over time as the TMS is used by the team to perform the task. Effective coordination and
interactions between team members, open access to knowledge, and demonstrations

of ability, all
provide the basis from which
perceptions of trustworthiness and so decisions to trust

are increased. In
contrast, poor coordination, low quality interaction or restricted access to knowledge may limit trust
development (Prichard & Ashleigh
, 2007). Therefore, we propose that in a reciprocal way, trust both
mobilizes and structures the TMS, and in turn the TMS structures and mobilizes trust.

In relation to our second aim, we have advanced the TM literature by proposing that perceptions of
tr
ustworthiness must be broadened rather than simply grounded in terms of trust in ability as is currently
concluded across TM research. In deciding whether or not to share knowledge with someone, or
evaluate their suitability to perform a particular role, t
eam members do not only consider whether or not
the other has the necessary expertise. Rather, we theorize that decisions
to trust others in these ways
are additionally grounded in perceptions of others’ trustworthiness in terms of whether they are working

for the interests of the team and act
ing

in ways that fit with the moral position of the group.



Managerial Implications

The propositions made in this paper reflect the importance of addressing the complexities of the TM
-
trust relationship in teams. It
points to the need to attend to the issues surrounding different factors
associated with trustworthiness and trust as a TMS develops. Specifically, we consider for the first time
the way that these act as antecedents of TM development as well as outcomes o
f a TMS in operation,
ultimately influencing successive encoding cycles and the future structure of the TMS. Understanding
the TM
-
trust relationship is important in organizational contexts as it allows for opportunities to improve
TMSs and so take advantag
e of storing and retrieving knowledge in teams. In turn, an effective TMS
enables teams to benefit from the cognitive savings that distributed knowledge
-
sharing systems can
offer. Therefore, knowledge about the implications of a lack of trust on the effect
ive use of distributed
26


knowledge enables managers to stay alert to interpersonal relations and take action when problems in
trust appear. Failure to address such issues could result in a team of experts failing to share their
knowledge for the good of the
team, and ultimately for the good of the organization.

Additionally, our analysis suggests that the promotion of trust in newly formed teams is likely to
promote TMS development. Consequently, time spent developing trust between team members when
the team

first forms could be an important focus for human resource management where those teams
are performing tasks in which a distributed knowledge sharing system is likely to offer advantages.
Although not directly linked to TM, a number of studies have shown
that trust promotion when a team is
first formed facilitates task performance (e.g.
Beranek, 2005;
Jarven
paa, Knoll & Leidner, 1998;
Prichard & Ashleigh, 2007). Ashleigh and Prichard (2011) propose that team
-
skills training may be an
effective mechanism fo
r promoting trust across a range of dimensions, and linked this to TMS
development in teams. To date, research on training to promote trust in teams is limited; therefore, in
addition to the propositions offered here, future research may wish to address fu
rther the role of training
to promote the TM
-
trust relationship.




Future Research

In this paper we have set out some specific hypotheses associated with our propositions which serve
as tangible examples of what may occur during TMS development. Clearly o
ther hypotheses are
possible relating to the different stages of the
encoding
cycle, the types of transactive processes
engaged in, the type of trust considered and the effect these variables have on TMSs. For example, one
important area of future research

within a trust domain could be to explore the transformation of trust
over time as the TM cycle continues and people become more specialized within their own knowledge
domain. Researchers such as McAllister (1996) and Lewicki and Bunker (1996) have propos
ed
transformational models of trust which suggest that the nature of trust transforms over time. A number
27


of such models have been proposed, each adopting slightly different frameworks (for review, see
Lewicki et al., 2006). These frameworks were initiall
y proposed to explain the articulation of trust
development across different types of relationships, e.g. intimate, personal and business, but also to
consider how the type and structure of trust relationships might change over time. Although, as in much
of the trust literature, a coherent and consistently agreed framework is yet to emerge, the idea that trust
changes over time may have implications for TMSs. For example, it raises questions such as; how do
the stages of trust identified in these transform
ational frameworks fit with different stages of the
encoding cycle, and do they impact differentially on how knowledge is shared
and accessed? A further
necessay focus of enquiry is to consider the differential effects of each factor of trustworthiness acr
oss
the stages of the TM encoding cycle. Mayer et al. (1995) suggest that integrity may be the most
important factor of trustworthiness in the early stages of a relationship, and therefore maybe more
salient in determining the decision to trust when initia
l claims about expertise are being made and
evaluated. The relative importance of ability and benevolence at different stages is less easy to predict
and likely to be dependent on issues such as the type of team (Jarnenpaa & Majchrzak, 2008) and the
nature

of the task being performed (Brandon & Hollingshead, 2004). While we have attempted to
integrate the trust and TMS literature and have provided a first theoretical framework from which to test
several hypotheses, we believe that questions relating to a co
-
evolutionary model of trust and TM and
the differential effects of the different factors of trustworthiness on TMSs should be the next phase of
extending this work. Future research will need to use both field and quasi
-
experimental studies that
would fram
e and test such questions.

Future research should also consider variables which may affect the TM
-
trust relationships we have
outlined in our analysis. One such variable is leadership, which may further complicate the relationship
between TM and trust by
influencing team dynamics. For example, leaders will adopt different
leadership styles (patterns of behavior) which they use to manage the team (Lee, Gillespie, Mann &
Wearing, 2010; Meyer, Ashleigh, Jones & George, 2007). These behaviors may for example i
nfluence
28


the degree of monitoring and control that managers hold over their subordinates (Ferrin, Bligh & Kohles,
2007). Some styles of leadership are likely to encourage more trusting relationships and feelings of
safety
and well being
in the team that wi
ll promote the sharing of knowledge

(Frazier, Johnson, Gavin,
Gooty & Snow, 2010)
. For example, Zhu, Avolio and Walumbwa (2009) propose that transformational
leaders facilitate a sense of psychological safety in their followers which makes it more likely t
hat they
will engage in knowledge sharing. Similarly, Chun, Litzky, Sosik, Bechtold and Godshalk (2010) have
argued that leaders who show high levels of emotional intelligence are more likely to build trusting
relationships with their followers. We therefo
re propose that styles of leadership that evoke trust will help
to mobilize knowledge sharing and promote memory specialization by affording higher levels of
autonomy to followers to develop expertise in specific components of the task. The literature on
l
eadership therefore signals that the relationship between TM and trust is influenced by the skills and
behavioral styles of the team’s leader and such nuances should be explored by future research.

The theorization of the TM
-
trust relationship may be furt
her moderated both by the type of task being
performed and the context in which the team is operating. For example
,

some tasks require lower levels
of cognitive interdependence and therefore limit reliance on distributed knowledge systems (Brandon &
Hollin
gshead, 2004
; Lewis & Herndon, 2011
). Under such circumstances, the TM
-
trust relationship will
be less important to team performance. The context in which teams work will also influence the
relevance of the TM
-
trust relationship
. F
or example, in virtual t
eams, the absence of social cues and
loosened control over team members challenges the development of trust (Peters & Karren, 2009). In
these contexts, where trust is harder to come by, our model of the TM
-
trust relationship would predict
that TM may be pa
rticularly difficult to develop and maintain. Future research should therefore consider
the theoretical model presented in this paper across a range of different types of tasks and contexts to
assess the extent of its applicability in contemporary teamwork
.

As has been detailed in each stage of the encoding cycle, there are complexities that are embedded
in the TM
-
trust relationship which we believe could be successfully explored through examining
29


experimental or quasi
-
experimental teams. Measures of trust
could be taken across several dimensions
to explore its relationship with the different stages of TMS development. For the sake of simplicity this
could commence with dyadic teams and then be extended to teams with more than two members. In
pursuing such l
ines of enquiry, researchers should think carefully about how to approach group
-
level
analysis of the role of trust in TMS development. Previous group
-
level research on TMSs has
considered individual team members’ ratings of each other, and then aggregated

individual scores to
arrive at group measures of TMS or team trust (Kanawattanachai & Yoo, 2007; Prichard & Ashleigh,
2007; Serva, Fuller & Mayer, 2005). Aggregation, however, risks losing the nuances of individual team
members’ relationships and interact
ions with one
another (De Jong & Dirks, 2010), and

Klein and
Kozlowski (2000) advise researchers to justify the need for aggregation and to select appropriate
statistical tools. For example, team member A may trust member B, but have limited trust in membe
r C.
The lack of trust in one person might have very specific effects on knowledge distribution in the team
which are not easily understood by reference to trust measured at the team level but which are revealed
by considering individual dyadic relationshi
ps. In progressing such work, we must not lose sight of the
complexity of trust. Given the interdependence of the different factors of perceived trustworthiness,
failure to consider benevolence and/or integrity may lead to a misguided faith that beliefs ab
out ability
are enough to ensure that a TMS will develop and then be utilized effectively. Even though the TM
literature is well established and reputable work in the field exists, we consider that further research is
needed to establish the relationship b
etween TM and trust across all of its factors. It is hoped that the
propositions presented here will provide a starting point for such work.


REFERENCES

Akgün, A. E., Byrne, J.B., Keskin, H.A., Lynn, G. S., & Imamoglu, S.Z. (2005). Knowledge networks in
n
ew product development projects: A transactive memory perspective.
Information & Management
,
42
: 1105

1120.

Ardichvili, A. Page, V., & Wentling, T. (2003).
Motivation and barriers to participation in virtual
knowledge
-
sharing communities of practice.
Jour
nal of Knowledge Management
,

7
, 64
-
77.



30


Ashleigh, M.J. & Prichard, J.S
.

(2011). Enhancing trust through training.
In D. Skinner & R. Searle (Eds.)

Trust
& Human Resource Management.

Edward Elgar Publishing.

Austin, J.R. (2003). Transactive memory in org
anizational groups: The effects of content, consensus,
specialization, and accuracy on group performance.
Journal of Applied Psychology, 88
, 866
-
878.

Beranek, P.M. (2005). A comparison of relational and trust training techniques for virtual team communicat
ion:
How much training is enough?
Proceedings of the 38
th

Hawaii International Conference on System
Sciences
.

Bradley, J., White, B. J., & Mennecke, B. E.
(
2003
)
. Teams and tasks: A temporal framework for the
effects of interpersonal interventions on team
performance.
Small Group Research
,
34: 353

387.

Brandon, D.P. & Hollingshead, A.B. (2004). Transactive memory systems in organizations: Matching
tasks, expertise and people.
Organization Science, 15
(6), 633
-
644.

Brodbeck, F. C., Kerschreiter, R., Mojzisch,

A. & Schulz
-
Hardt, S. (2007).
Improving group decision
making under conditions of distributed knowledge: The information asymmetries model.
Academy of
Management Review
,
32
, 459
-
479.

Cattell, R. B. (1965). The scientific analysis of personality. Baltimore

M.D: Penguin.

Chatman & Flynn (2001).The influence of demographic heterogeneity on the emergence and
consequences of cooperative norms in work teams.
Academy of Management Journal, 44,

956
-
974.

Chun, J.U., Litzky, B.E. Sosik, J.J., Bechtold D.C., & Godsha
lk, V.M. (2010). Emotional intelligence and
trust in formal mentoring programmes.
Group &

Organization Management, 35,

421
-
455.

Cummings, L.L., & Bromiley, P. (1996). The Organizational trust inventory (OTI): Development and
Validation. In: R. .M. Kramer
, & R. R. Tyler
,

(Eds.),
Trust in organizations: Frontiers of theory and
research
, 302
-
330. London: Sage Publications.

Curşeu, P.L. (2006). Emergent states in virtual teams. A complex adaptive systems perspective.
Journal
of Information Technology
, 21, 4, 249
-
261.

Curşeu, P.L. & S.G.L. Schruijer (2010). Does conflict shatters trust or does trust obliterate conflict?
Revisi
ting the relationship between team diversity, conflict and trust.
Group Dynamics: Theory,
Research and Practice
,

14
,

66
-
79.

De Jong, B. A., & Elfring,

T. (2010).
How Does Trust Affect the Performance of Ongoing Teams? The
Mediating Role of Reflexivity, Mon
itoring and Effort.
Academy of Management Journal,

53(3), 535

549.

De Jong, B. A., & Dirks, K. T. (2010, January 28
-
29).
Control and Trust within Teams: Joint Implications
of Mean and Dispersion Levels
. Paper presented at the 5th EIASM Workshop on "Trust w
ithin and
between Organizations", Madrid, Spain.

31


Dirks, K.T., & Ferrin D.L. (2001). The role of trust in organizational settings.
Organization Science
,
12
(4),
450
-
467.

Edmondson, A.C. (2003). Speaking up in the operating room: How team leaders promote lear
ning in
interdisciplinary action teams.
Journal of Management Studies,

40
(6),

1419
-
1452.

Enzle, M. E., & Anderson, S. C. (1993).
Surveillant intentions and intrinsic motivation.
Journal of
Personality and Social Psychology,
64
, 257

266.

Faraj, S., & Sproul
l, L.

(2000). Coordinating expertise in software development teams.
Management
Science
,
46
(12), 1554
-
1568.

Ferrin, D.L., Bligh, M.C., & Kohles, J.C. (2007). Can I trust you to trust me?: A theory of trust, monitoring
and cooperation in interpersonal and in
tergroup relationships.
Group &

Organization Management,
32
, 465
-
499.

Ferrin, D. L., Bligh, M. C., & Kohles, J. C. (2008). It takes two to tango: An interdependence analysis of
the spiraling of perceived trustworthiness and cooperation in interpersonal and

intergroup
relati
onships. Organizational Behavio
r and Human Decision Processes, 107(2), 161
-
178.

Frazier, M. L., Johnson, P. D., Gavin, M., Gooty, J. & Snow, D. B. (2010). Organizational justice,
trustworthiness, and trust: A multifoci examination.
Group
& Organization Management,

35: 39
-
76.

Gill, H.. Boies, K., Finegan, J.E. & McNally, J. (2005). Antecedents of trust: Establishing a boundary
condition for the relation between propensity to trust and intention to trust.
Journal of Business and
Psychology,

19,

287
-
302.

Hogg, M.A., & Terry, D.J. (2000) Social identity and self
-
categorization processes in organizational
contexts.
Academy of Management Review
,
25
, 121
-
140.

Hollingshead, A. B. (1998
). Communication, learning and retrieval in Transactive memory
systems.
Journal of Experimental Social Psychology 34
, 423
-
442.

Hollingshead, A.B. (2000). Perceptions of expertise and transactive memory in work relationships. In:
(Eds),
Group Processes & Intergroup Relations

3
(3), 257
-
300.

Jarvenpaa, S.L., & Majchrzak
, A. (2008). Knowledge collaboration among professional protecting
national security: Role of transactive memories in ego
-
centred knowledge networks,
Organization
Science 19
(2), 260
-
276.

Jarvenpaa, S.L., Knoll K., & Leidner, D. E. (1998). Is anybody out th
ere? Antecedents of trust in global
virtual teams.
Journal of Management Information Systems,

14
(4), 29
-
64.

Kanawattanachai, P, & Yoo, Y. (2007). The impact of knowledge coordination on virtual team
performance over time.
MIS Quarterly, 31
, 783
-
808.

Kim W
.C., & Mauborgne, R. (1998). Procedural justice, strategic decision making, and the knowledge
economy.
Strategic Management Journal

19
(4), 323
-
338.

32


Klein, K. J., & Kozlowski, S. W. J. (2000).
From Micro to Meso: Critical Steps in Conceptualizing and
Conduc
ting Multilevel Research. Organizational Research Methods, 3(3), 211
-
236.

Koskinen, K.U., Pihlanto, P., & Vanharanta, H. (2003).
Tacit knowledge acquisition and sharing in a
project work context.
International Journal of Project Management, 2,

281
-
290.

Lan
gfred, C.W. (2004). Too much of a good thing? Negative effects of high trust and individual
autonomy in self
-
managing teams.
Journal of Management Review, 47,

385
-
399.

Lee, P., Gillespie, N., Mann, L., & Wearing, A. (2010).
Leadership and trust: Their effe
ct on knowledge
sharing and team performance.
Management Learning, 41
(4), 473
-
491.

Levin, D.L., Whit
e
ner, E. M., & Cross, R. (2006). Perceived trustworthiness of Knowledge Sources: The
moderating impact of relationship length.

Journal of Applied
P
sychology
, 91
(5), 1163
-
1171.

Lewicki, R. J., & Bunker, B. B. (1996). Developing and Maintaining Trust in Work Relationships. In R. M.
Kramer & T. R. Tyler (Eds.),
Trust in Organizations: Frontiers of Theory and Research

(pp. 114
-
139).
Thousand Oaks, CA: Sage.

Lewic
ki, R. J., Tomlinson, E.C., & Gillespie, N. (2006). Models on interpersonal trust development:
Theoretical approaches, empirical evidence and future directions.
Journal of Management
,
32
(6),
991
-
1002.

Lewis, K. (2003). Measuring transactive memory systems
in the field: Scale development and validation.
Journal of Applied Psychology, 88
, 587
-
604.

Lewis, K. (2004). Knowledge and performance in knowledge
-
worker teams: A longitudinal study of
transactive memory systems.
Management Science, 50
, 1519
-
1533.

Lewis,

K. & Herndon, B. (2011). Transactive memory systems: Current issues and future research
directions.
Organization Science, 22
(5), 1254
-
1265.

Lewis, D.J., & Weige
rt, A. (1985).
Trust as a social reality.
Social Forces, 63
(4), 967
-
985.

Li, X., Hess, T. J., &

Valacich, J. S. (2008).
Why do we trust new technology? A study of initial trust
formation with organizational information systems.
Journal of Strategic Information Systems
,
17
, 39
-
71.

Liang, D.W., Moreland, R., & Argote, L. (1995). Group versus individua
l training and group performance:
The mediating role of transactive memory.
Personality and Social Psychology Bulletin, 21
, 384
-
393.

Madhavan, R., & Grover, R. (1998). From embedded knowledge to embodied knowledge: New product
development as knowledge mana
gement.
Journal of Marketing, 62
, 1
-
12.

Mayer, R.C.,
R.C
.
,
Davis, J.R. & Schoorman, F.D. (1995). An integrative model of organizational trust.
Academy of Management Journal

20
(3), 709
-
737.

McAllister, D.J. (1995). Affect and cognition
-
based trust as found
ations for interpersonal cooperation in
organizations.
Academy of Management Review
,
38
, 24
-
59.

33


McEvily, B. Perrone, V. & Zaheer
, A. (2003). Trust as an organiz
ing principle.
Organization Science,
14
(1), 91
-
103.

McKnight, H.D., Cummings, L. L., & Chervany,

N. L. (1998).
Initial trust formation in new organizational
relationships.
The Academy of Management Review
,
23 (3)
, 473
-
490.

Mesmer
-
Magnus, J.R., & DeChurch, L.A. (2009). Information sharing and team performance: A meta
-
analysis.
Journal of Applied Psych
ology,

94, 2, 535
-
546.

Meyer, E., Ashleigh, M., Jones, G. & George, J. (2007).
Contemporary Management, European Edition.

Maidenhead, UK, McGraw Hill.

Meyerson, D., Weick, K. E., & Kramer, R. M. (1996).
Swift trust and temporary groups. In R.M. Kramer
&
T.T. Tyler (Eds.),
Trust in organizations: Frontiers of Theory and Research
. London: Sage
Publications.

Mishra A.K. (1996).
Organizational

responses

to

crisis
:
the

centrality

of

trust
. In

R Kramer, T Tyler (Eds.),
Trust in Organizations: Frontiers of Theo
ry and Research
,
261
-

287. Thousand oaks: Sage.

Mohammed, S., & Dumville, B. C. (2001). Team mental models in a team knowledge framework:
expanding theory and measurement across disciplinary boundaries.
Journal of Organizational
Behavior, 22
, 89
-
106.

More
land, R. L., & Myaskovsky, L. (2000). Exploring the performance benefits of group training:
Transactive memory or improved communication?
Organizational Behavior and Human Decision
Processes, 82,
117
-
133.

Nelson, K, M., & Cooprider, J. G. (1996).The contri
bution of shared knowledge to IS group performance.
MIS Quarterly
,

20
(4), 409
-
432.

Nonaka, I., & Krogh, G. (2009). Tacit knowledge and knowledge conversion: Controversy and
advancement in organizational knowledge creation theory.
Organization Science, 20
(
3), 635
-
652.

Peltokorpi, V. (2008).Transactive Memory Systems.
Review of General Psychology, 12,

378
-
394.

Peters L. & Karren, R.J. (2009).
An examination of the roles of trust and functional diversity on virtual
team performance ratings.
Group &

Organizati
on Management, 34
, 479
-
504.

Politis, J. D. (2003). The connection between trust and knowledge management: What are its
implications for team performance.
Journal of Knowledge Management,

7(5), 55
-
66.

Prichard, J. S., & Ashleigh, M. J. (2007). The effects o
f team
-
skills training on transactive memory and
performance.
Small Group Research, 38
(6)
,

696
-
726.

Prichard, J.S., Bizo, L.A., Stratford, B.J. (2011). Evaluating the effects of team
-
skills training on
subjective workload,
Learning and Instruction

(2010) d
oi:10.1016/j.learninstruc.2010.06.003.

Rau, D. (2005). The influence of relationship conflict and trust on the TM: Performance relation in top
management teams.
Small Group Research 36
, 746
-
771.

34


Rulke D.L., & Rau, D. (2000).
Investigating the encoding proc
ess of transactive memory development in
group training
. Group &

Organization Management
,

25
, 373
-
396.

Schoorman, F.D,. Mayer, R.C. & Davis, J.H. (2007). An integrative model of organizational trust: Past,
present, and future.
Academy of Management Review
,

32
(2), 344
-
354.

Serva, M.A., Fuller M.A., & Mayer, R.C. (2005). The reciprocal nature of trust: a longitudinal study of
interacting teams. J
ournal of Organizational Behavior
,
26
(6), 625
-
648.

Sharkie, R. (2005). Perceptions of trust as a determinant of th
e level of knowledge sharing.
International
Journal of Knowledge, Culture and Change Management
,
4
, 1795

1801

Sitkin, S.B., (1995). On the positive effect of legalization on trust. In R. Bies, B. Sheppard, & R. Lewicki,
(Eds.),
Research on Negotiations in
Organizations
, 185

217.
Greenwich: JAI Press.

Spreitzer, G., Sutcliffe K., Dutton, J., Sonenshein, S. & Grant, A.M. (2005).
A Socially Embedded Model
of Thriving at Work.
Organization Science, 16
(5), 537

549.

Staples, S. S., & Webster, J. (2008). Explori
ng the effects of trust, task interdependence and virtualness
on knowledge sharing in teams.
I
nformation Systems Journal,
18(6), 617
-
640.

Stasser, G. Stewart, D.D. & Wittenbaum, G. M.(1995).
Expert roles and information exchange during
discussion: The impo
rtance of knowing who knows what.
Journal of Experimental Social Psychology,
31
, 244
-
265.

Wegner, D. M. (1986). Transactive memory: A contemporary analysis of the group mind. In M.B. Mullen
& G. Goethals (Eds.),
Theories of Group Behavior,

185
-
205. New Yo
rk: Springer
-
Verlag.

Wegner, D. M. (1995). A computer network model of human transactive memory.
Social Cognition, 13
,
319
-
339.

Wegner, D. M., Giuliano, T., & Hertel. P. (1985). Cognitive interdependence in close relationships. W. J.
Ickes, (Ed.)
Compatible and Incompatible Relationships
.
New York: Springer
-
Verlag.

Zand, D. E. (1972). Trust and Managerial Problem Solving.
Administrative Science Quarterly
, 17(2),
229
-
239.

Zhu, W., Avolio, B.J., & Walumbwa F.O. (2009). Moderating role of follower cha
racteristics with
transformational leadership and follow work engagement.
Group &

Organization Management, 34
,
590
-
619.

35



Members claim
expertise or
declare a lack of
expertise

Coordination of
responsibilities
based on
expertise
distribution
identified


Task
performan
ce
feeds into future
evaluations of
expertise

Expertise
evaluated to
establish
knowledge
distribution and
gaps

Stage 2:

Factors of
trustworthiness and trust affect

evaluations of claims of
expertise

made in stage 1

Stage 1:

Factors of
trustworthiness an
d trust are

antecedent
s

for the declaration
of knowledge

Stage 3
:

Factors of
trustworthiness and trust

influence

coordination
processes and the degree of
memory specialization

Stage 4
:

Factors of
trustworthiness and trust
are
outcomes of TMS and
in
fluenc
e

successive encoding cycles



Trust

based on

perceptions of

trustworthiness
:

ability, benevolence
and i
ntegrity


Figure 1
: Summary of
the
relationship between transactive memory, factors of perceived trustworthiness and trust
. The central circle sh
ows the
TM encoding cycle adapted from Rulke and Rau (2000)