Content uploaded by Ephraim McLean
Author content
All content in this area was uploaded by Ephraim McLean on Oct 06, 2014
Content may be subject to copyright.
EXPERTISE INTEGRATION AND CREATIVITY IN INFORMATION SYSTEMS DEVELOPMENT 13
Journal of Management Information Systems / Summer 2005, Vol. 22, No. 1, pp. 13–43.
© 2005 M.E. Sharpe, Inc.
0742–1222 / 2005 $9.50 + 0.00.
Expertise Integration and Creativity in
Information Systems Development
AMRIT TIWANA AND EPHRAIM R. MCLEAN
AMRIT TIWANA is an Assistant Professor at Iowa State University in Ames, and was
an Assistant Professor at Emory University in Atlanta, when this work was com-
pleted. His Ph.D. is in Management Information Systems from Georgia State
University’s Robinson College of Business. His research focuses on knowledge man-
agement in information systems development. His work has been published or is
forthcoming in California Management Review, IEEE Transactions on Engineering
Management, IEEE Software, IEEE Internet Computing, Communications of the ACM,
Decision Support Systems, Information Systems Journal, Journal of Knowledge Man-
agement, International Conference on Information Systems, Hawaii International
Conference on System Sciences, and others. He is also the author of The Knowledge
Management Toolkit (2d ed., Prentice Hall, 2002).
EPHRAIM R. MCLEAN is a Regents’ Professor and George E. Smith Eminent Scholar’s
Chair in Information Systems in the Robinson College of Business at Georgia State
University in Atlanta. Prior to coming to Georgia State University in 1987, he was on
the faculty of the University of California, Los Angeles (UCLA) for 18 years. Dr.
McLean’s research focuses on the management of information services, the value of
IS investments, and career issues for IS professionals. He has published over 125
papers in such journals as Information Systems Research, Journal of Management
Information Systems, MIS Quarterly, Management Science, Communications of the
ACM, DATABASE, Harvard Business Review, Sloan Management Review, and others;
and his coauthored book, Information Technology for Management, now in its fourth
edition, is currently the second-largest selling IS textbook in the world. Dr. McLean
earned his B.M.E. in mechanical engineering from Cornell University and his S.M.
and Ph.D. from the Sloan School of Management at the Massachusetts Institute of
Technology (MIT). He is also the Executive Director of the Association for Informa-
tion Systems (AIS) and in 1999 was made a Fellow of the AIS.
ABSTRACT: This paper addresses the understudied issue of how individually held ex-
pertise in information systems development (ISD) teams results in creativity at the
team level during the development process. We develop the idea that team creativity
results primarily from integration of individually held expertise of team members at the
team level. We further propose the quality of intrateam relationships and knowledge
complementarities that align the work of individual team members at the project level
influence creativity primarily through the process of expertise integration. We use data
from a field study of 142 participants in 42 ISD projects to test the proposed model.
The paper makes three new contributions to the IS literature. Its key contribution lies
in developing an expertise integration view of team creativity. We demonstrate the
centrality of integrating individually held tacit and explicit knowledge about the prob-
lem domain and the technology at the team level in achieving team creativity. The use
of a process-focused conceptualization of team creativity is especially noteworthy here.
01 tiwana.pmd 6/15/2005, 11:21 PM13
14 AMRIT TIWANA AND EPHRAIM R. MCLEAN
The second contribution of the paper lies in conceptually developing and
operationalizing the concept of expertise integration, a mechanism by which individu-
ally held knowledge is integratively applied at the project level. Although the impor-
tance of knowledge in the ISD process is widely recognized in prior research, this is
the first study to develop the concept in a operationally meaningful way. The third key
contribution lies in showing that the compositional and relational attributes of ISD
project teams—diverse specialized knowledge in a team, the quality of intrateam work-
ing relationships, and members’ cross-domain absorptive capacity—do not engender
creativity by themselves; they do so primarily because they enhance integration of
individual knowledge at the project level. We offer empirical evidence for such full
mediation. These findings have important theoretical and practical implications, which
are discussed in the paper.
KEY WORDS AND PHRASES: absorptive capacity, creativity, expertise integration, in-
formation systems development, information systems innovation, integration, knowl-
edge integration, knowledge management, knowledge transfer, software development.
DEVELOPMENT OF INFORMATION SYSTEMS (IS) is a creative effort that involves the
expertise, insights, and skills of many individuals. As organizations encounter the
need to develop systems for novel business applications (e.g., knowledge manage-
ment, peer-to-peer collaboration) and new problem domains (e.g., reverse logistics in
supply chains), the need for creativity in the information systems development (ISD)
process is increasingly recognized in practice. However, the creativity construct it-
self remains narrowly studied in the IS literature. A review of over two decades of
literature (1981–2003) reveals three themes in IS studies on creativity. First, studies
have focused primarily on creativity as a causal outcome of IS use. Examples include
studies of how decision support systems enhance individual creativity (e.g., [42]),
how creativity support systems (CSS) influence the quality and quantity of individual
creative output (e.g., [43, 73]), and how the use of software tools stimulates indi-
vidual creativity [23, 41, 49, 58, 59]. No prior study has examined creativity during
the ISD process. Second, the unit of analysis used to examine creativity has primarily
been individuals, whereas most ISD in contemporary organizations is carried out in
project teams. Third, IS researchers have operationalized creativity as a novel and
useful outcome, as judged by experts. However, creativity researchers in other disci-
plines have cautioned that this reductionist approach is biased toward successful out-
comes and prone to retrospection bias [22, 66]. Thus, while creativity is recognized
as an important element of ISD, gaps in the existing literature prevent us from an-
swering the basic question of how expertise in an ISD team translates into creative
processes. In the present study, we address two research questions that follow from
this overarching research problem.
01 tiwana.pmd 6/15/2005, 11:21 PM14
EXPERTISE INTEGRATION AND CREATIVITY IN INFORMATION SYSTEMS DEVELOPMENT 15
1. How does the expertise of individual team members influence an ISD project
team’s creativity?
2. How do the compositional attributes of an ISD team influence its creativity?
We theoretically develop the idea that an ISD team’s creativity is predicted by the
extent to which its members integrate their specialized expertise to jointly develop
project concepts, designs, and solutions. We refer to such recombination of indi-
vidual expertise at the project level as expertise integration. We then focus on some
antecedents of expertise integration that are rooted in the compositional attributes of
the team: heterogeneity of expertise, the quality of the relationships among team
members, and their ability to interrelate with the expertise of their peers outside of
their own domain. Their influence on creativity is fully mediated by expertise inte-
gration. The model is empirically tested through a field study of 142 participants in
42 ISD project teams and their senior management project stakeholders.
The study makes three novel contributions to the ISD literature. First, it conceptu-
alizes the perspective that individually held expertise influences creativity in the ISD
process primarily through the process of expertise integration at the team level. This
is an important contribution, because it highlights the central role of expertise inte-
gration in facilitating team creativity—a relationship that has not previously been
conceptually developed or tested. This has pervasive implications for ISD, because it
shifts attention away from knowledge transfer (as implicitly assumed in ISD activi-
ties such as requirements specification and models such as the waterfall approach)
toward its integration at the project level. The second noteworthy contribution of this
paper is in showing how integration of individually held expertise at the team level
enhances team creativity in the ISD process. With the sole exception of Cooper’s
[15] descriptive study, no prior study has examined the role of expertise integration
in enhancing creativity in ISD. It is noteworthy that although the importance of di-
verse knowledge in the ISD process is well recognized in prior IS research, this is the
first study to actually measure the concept of expertise integration and empirically
test its relationship with team creativity. The third important contribution is demon-
strating that the compositional attributes of the project team—heterogeneity in team
members’ expertise, the quality of the working relationships within the team, and
their collective absorptive capacity—influence the extent to which its members can
integrate their diverse expertise bases in formulating a coherent set of ideas for the
project solution. This lends new insights into managing ISD projects, which contin-
ues to be a dominant challenge in contemporary organizations [50, 71], especially
aligning ISD projects with their business objectives [39, 68]. Although it was not the
focus of this study, we also demonstrate that higher levels of creativity at the team
level are associated with higher levels of project success. This finding complements
prior work on improving ISD project outcomes by using tools for enhancing ISD
productivity [18, 62], project control mechanisms [47], and utilization of a variety of
methodological variations [34]. In summary, the study contributes new insights into
a key mechanism—the process of integrating individual knowledge at the project
01 tiwana.pmd 6/15/2005, 11:21 PM15
16 AMRIT TIWANA AND EPHRAIM R. MCLEAN
level—through which individually held expertise leads to creativity during the ISD
process.
Theoretical Development
IN THIS SECTION, WE FIRST EXAMINE the theoretical domain of creativity in ISD. We
then develop the idea that creativity at the team level emerges from integrating indi-
vidually held expertise at the project level. Such integration, in turn, is facilitated by
the compositional antecedent characteristics of the project team’s knowledge and
relational attributes.
Team Creativity
Team creativity is defined as the degree to which a project team’s processes are novel
in the context of the project’s objectives [22]. Creative processes can exist both at the
individual level and at the group level. Since researchers have only recently paid
attention to the differences between individual and team creativity, it is worthwhile
considering how the two differ. The noteworthy difference is that creativity at the
team level is an inherently social process [53]. Individual contributions are a part of,
but not the entirety of, team creativity.1 Team creativity emerges from an improvisa-
tional process where individual team members collaboratively build on and interre-
late their ideas with the perspectives and unique skills of other individuals in the
project team, so that the joint activities of individual team members create a collec-
tive system of creative actions. In addition, there are occasions where one individual’s
actions spark collective actions that are creative in nature. Thus team creativity is
inherently a social process that builds on and incorporates individual creative pro-
cesses at the project level.
ISD is an inherently creative process because it involves generation and evaluation
of new ideas, designs, solutions, and artifacts (for a review, see [51]). The systems
development life cycle involves translating an abstract business idea into project re-
quirements, which are then used to create project concepts and system specifications,
and eventually the functionality and features in the software code. Successful ISD
thus depends on a team’s developing a preliminary idea beyond its embryonic state
by drawing on several interdependent bases of expertise. There is rarely “one right
design” for an ISD problem, because there is often more than one possible solution to
the same end [55]. Depending on how creative the process is, a team might come up
with one of many possible solutions to the same problem. Thus a creative develop-
ment process is likely to explore multiple possible solutions.
The importance of creativity in the ISD process has only recently been explicitly
recognized [15], and the construct itself has received little direct attention in IS re-
search. The collective nature of creativity has been largely unexplored [53].2 The
prevalent conceptualization of creativity as a novel and useful outcome further blurs
the distinction between creative processes and outcomes.3 An approach that assesses
01 tiwana.pmd 6/15/2005, 11:21 PM16
EXPERTISE INTEGRATION AND CREATIVITY IN INFORMATION SYSTEMS DEVELOPMENT 17
the degree to which a team’s processes are creative recognizes that creative processes
are a necessary but insufficient condition for creative outcomes [22], and is thus more
appropriate for assessing creativity in ISD teams [61]. It also ensures that creative
processes are conceptually separable from their outcomes. Team creativity, in turn,
influences project outcomes—an assumption that is later empirically confirmed in
this study.
Expertise Integration
At the core of the knowledge-based view of organizing is the conception of organiza-
tional entities such as project teams as vehicles for integrating tacit and explicit knowl-
edge that is distributed among many individuals [27]. While expertise is “owned” at
the individual level, it is necessary to integrate specialized, individually held exper-
tise into collective (project) knowledge for a project to benefit from it [52, 61]. Build-
ing on Grant’s [27] initial conceptualization and recent team-level extensions of Grant’s
definition [2, 52], we define expertise integration as the coordinated application of
individually held specialist expertise in the accomplishment of tasks at the project
level.4 Expertise is integrated when at least one piece of knowledge from one indi-
vidual is used together with expertise from another team member to accomplish a
project task. Thus, integration of the understanding and expertise of different indi-
viduals at the project level renders it usable for creating the system.
Central to the process of integrating expertise is the conversion of knowledge con-
sisting of socially derived scripts, interpretations, and goals of individual team mem-
bers into a coherent software system. Much of the prior research on knowledge
management has focused on knowledge transfer and knowledge sharing. Little atten-
tion, however, has been paid to expertise integration, which refers to the process by
which existing individually held expertise is brought to bear at the project level [2].
Although expertise integration is one mechanism for applying the knowledge of many
individuals in a team to the ISD process, in theory this end can also be achieved
through knowledge transfer or knowledge sharing. It is therefore useful to distin-
guish expertise integration from knowledge transfer and sharing.
For simplicity of illustration, consider a two-person setting where person A and
person B each possess one piece of unique knowledge K(a) and K(b) before the trans-
fer/sharing/integration process. Knowledge transfer refers to transmission of knowl-
edge from one individual to another [2]. Ideally, at the end of such a knowledge
transfer process, the transferee should possess the transferred knowledge K(a) of the
transferor in its entirety [17]. Therefore, the widely used waterfall model of ISD
implicitly assumes that such knowledge transfer is possible during the initial require-
ments determination phase of systems development. However, the presence of tacit
components of knowledge makes knowledge transfer—while theoretically feasible—
a less viable mechanism for applying members’ knowledge to the ISD process. The
primary reason is that knowledge transfer is an inherently slower and inefficient pro-
cess [64]. As our results later illustrate, the typical project in this study lasted only
about six months, leaving little time for knowledge transfer to be viable. Knowledge
01 tiwana.pmd 6/15/2005, 11:21 PM17
18 AMRIT TIWANA AND EPHRAIM R. MCLEAN
sharing is a more limited instance of knowledge transfer: It involves revealing the
presence of pertinent knowledge without necessarily transmitting it in its entirety
(i.e., B’s gaining only a subset of knowledge K(a) from A via knowledge sharing).
The viability of knowledge sharing as an approach for knowledge application in soft-
ware projects is problematic for two reasons. First, the tacit elements of knowledge,
such as a customer’s understanding of his or her needs, are difficult to fully articulate
up front in the form of formal project requirements or system specifications. Second,
knowledge sharing does not ensure that the knowledge shared by individual project
participants will be utilized during the ISD process. Moreover, technical change or
requirements volatility might render some knowledge irrelevant by the time it is trans-
ferred or after it is shared.
Effective teamwork emerges from new knowledge that results from interactions
among specialists in a team, not simply from individual gains in knowledge by indi-
vidual team members [52]. In other words, individuals in the team must integrate the
knowledge that is shared at the project level to realize its value. When considerable
tacit knowledge is involved, coordinated application of the knowledge held by many
individuals to the project is more feasible through expertise integration, because it
does not incur the time and costs needed for team members to teach each other their
jobs (i.e., full-fledged knowledge transfer). Expertise integration involves develop-
ment of new project-level knowledge based on the insights contributed by many indi-
viduals. Unlike knowledge sharing or knowledge transfer, expertise integration builds
on, but goes beyond, specialized domain expertise. It requires both sharing of indi-
vidually held knowledge with the rest of the team and utilization of the shared knowl-
edge in the context of the project. Expertise integration therefore cannot be reduced
to knowledge sharing, although it involves some degree of iterative knowledge shar-
ing. The outcome of expertise integration is new project-level knowledge that syn-
thesizes insights from the multiple thought worlds of the team members [21].
Project-level knowledge that results from the process of integration is idiosyncratic
to the project and is held collectively at the team level [52]. Expertise integration thus
creates a negotiated set of beliefs among the project team members. Such integration
creates a “mutual equivalence structure” [72] wherein each team member recognizes
that the ISD process consists of his own and other team members’ interdependent
expertise, and constructs individual actions to be relevant to project activities.
Expertise Integration in ISD Project Teams
Fully understanding the problem that the intended system must solve is often one of
the most challenging aspects of ISD [32]. Developing a system that addresses the
problem domain requires two types of knowledge: Technical knowledge and knowl-
edge about the application problem domain [57], much of which is dispersed among
different project stakeholders such as analysts, domain experts, programmers, and
potential users [16, 32]. Such technical and application domain knowledge must be
integrated at the project level in formulating project concepts and solutions. Some of
it—even when available within the team—is not readily available in an explicit form
01 tiwana.pmd 6/15/2005, 11:21 PM18
EXPERTISE INTEGRATION AND CREATIVITY IN INFORMATION SYSTEMS DEVELOPMENT 19
(such as requirements documents) or captured in formal specifications [9, 69].5 More-
over, ISD largely consists of heuristic tasks—that is, tasks that do not have clear and
readily identifiable paths to the solution [15]. This makes it difficult to fully transfer
knowledge through requirements documents and formal specifications.6 Unless this
knowledge is integrated during the ISD process, it is unlikely that artifacts of the ISD
process (e.g., requirements documents, specifications, features, use cases, code, and
documentation) will accurately capture the capabilities that users need from the sys-
tem. For example, information that software architects and programmers need in the
ISD process is often not available to them through documents, although analysts and
end users may have this knowledge [38]. When integration of team members’ exper-
tise at the project level is poor, the team might build its ignorance—unstated require-
ments, evolving user needs, unrecognized constraints, and incomplete understanding
of the problem domain—into the design of the system. Thus some of the considerable
risk involved in the ISD process stems from the risk that the delivered solution simply
fails to meet changing business needs [6, 7].
Expertise Integration and Creativity
Next, consider how expertise integration influences creativity at the team level. Highest
levels of team creativity are a result of different cognitive structures coming together
[46]. Team creativity results from finding novel associations and linkages among the
diverse ideas, perspectives, and domain expertise that individual team members hold
[4, 15, 30, 63]. Individuals in a team often bring different ideas, perspectives, and
expertise to the project. Access to a variety of alternatives, example solutions, and
ideas can potentially lead to higher team creativity [4]. But realizing this potential for
creativity requires that such ideas be relevant to the project. This requires that indi-
vidually held expertise be applied to project activities with an appreciation of the
project context, its business needs, and constraints.
ISD projects must draw on and integrate the contributions of various members into
a coherent solution [14]. Integration of individually held expertise at the team level
provides a mechanism for enhancing team creativity, because it leads team members
to access, explore, and use diverse information from related knowledge domains as-
sociated with the project. Individual team members, however, often start out with
their own partial mental models about the design problem and possible solutions,
which are biased by their prior experience in other ISD projects [55]. Before connec-
tions among diverse ideas and viewpoints can be explored, individual team members
must first overcome their preexisting biases and stereotypes that are deeply rooted in
their functional identities and backgrounds [74]. Team members’ exposure to differ-
ent alternatives, approaches, and ideas can trigger a generation of divergent solutions
[53]. As they engage in integrating their understanding of the project with other team
members, the team can arrive at an agreeable shared vision of the system’s structure,
function, interdependencies, and needed capabilities. This common understanding
brings various team members closer to a common understanding of the requirements
that the system must address [32]. Expertise integration thus creates a shared under-
01 tiwana.pmd 6/15/2005, 11:21 PM19
20 AMRIT TIWANA AND EPHRAIM R. MCLEAN
standing about the project within which novel associations among individually held
diverse expertise bases can be created [22, 44]. This shared understanding allows the
ISD team to remain cognizant of various technical, operational, and economic con-
straints, which might not be known in their entirety to any single individual in the
team. This facilitates consideration of design and implementation problems from less
obvious perspectives, because the ISD process then draws on the expertise of many
individuals in generating and evaluating potential design solutions. Expertise inte-
gration thus facilitates experimentation with novel associations among the expertise,
viewpoints, potential solutions, and perspectives held by individual team members,
thereby enhancing the ISD team’s creativity. Integration of individually held exper-
tise at the team level is therefore necessary for creative processes to emerge in the
ISD process. This leads to our first hypothesis.
H1: Expertise integration is positively and directly related to creativity in ISD
teams.
Compositional Antecedents of Team Creativity. Clearly, one of the challenges facing
managers is that of organizing a team at the outset of a project that is more likely to be
creative during the ISD process. This is a question concerning team design, which is
defined as the clustering of individuals and their relationships into a project team, and
the knowledge complementarities that align their work at the project level. The most
diverse collection of expertise bases in a given team is likely to result in potentially
the most creative teams, but realizing creativity in a diverse, interdependent group of
individuals in a team is another issue. This dilemma poses a problem of organizing a
team that will likely be creative, rather than merely have the potential to be creative.
Given the central position of the perspective that team creativity in ISD projects
emerges through the process of integrating individually held expertise at the team
level, we chose to focus on a subset of antecedent variables that collectively specify
some team-level compositional attributes that influence expertise integration at the
team level. Although several factors have been linked to creativity, three important
factors are directly relevant to this view—the diversity of knowledge available in a
team, the working relationships among team members, and the extent to which the
team members can interrelate with each others’ expertise. Normally, such expertise
integration is difficult to achieve in cross-functional project teams, because people
from different functional areas hold biases and stereotypes toward one another, which
need to be adapted. The most profound challenges to integrating knowledge come
from the ability of individuals to utilize the expertise of their peer team members in
their collective project activities and from the quality of relationships among them
[64]. We therefore focus on three key antecedents of ISD team creativity—expertise
heterogeneity, relational capital, and absorptive capacity.7 These variables capture
the key intra team properties of the content of individually held expertise and the
relationships through which that expertise is applied to project activities. The con-
ceptual logic that we develop is that these variables influence team creativity by fa-
cilitating expertise integration. Clearly, this set of factors is a subset of all possible
factors that can influence expertise integration, as acknowledged later in the paper.
01 tiwana.pmd 6/15/2005, 11:21 PM20
EXPERTISE INTEGRATION AND CREATIVITY IN INFORMATION SYSTEMS DEVELOPMENT 21
Expertise Heterogeneity
ISD projects must draw on knowledge from multiple technical and functional do-
mains [16, 70]. One mechanism for gaining direct access to such diverse expertise is
to have team members with heterogeneous expertise, who will bring to the project
different sets of skills, perspectives, and knowledge that are less likely to be available
to a homogenous team [5, 10]. We define expertise heterogeneity as the diversity in
the expertise possessed by the members of a project team. As the number of func-
tional domains represented by individual members increases, so does the variety of
ideas and the range of possible linkages and associations among those ideas. Indi-
vidual team members can then borrow ideas and concepts from outside their own
domains, often drawing different implications from the same ideas. Recombination
of ideas that are old in a specific knowledge domain but new to the team can lead to
creative new ways of formulating and conceptualizing project ideas. In contrast, indi-
viduals with similar backgrounds and experiences tend to view things similarly, thereby
missing opportunities for exploring novel design ideas. Members of a heterogeneous
team also provide access to a broader set of external networks [5]. This allows the
team to draw on a broader set of external sources of information and knowledge
during the ISD process.8 Therefore, a team with a heterogeneous expertise base is
likely be better at exploring design decisions at depth, pose more diverse alternatives,
and challenge assumptions that might be readily accepted in more homogenous teams.
However, too much expertise heterogeneity can be a mixed blessing: while heteroge-
neity brings a more varied pool of expertise for potential recombination, the difficul-
ties in reconciling diverse interpretations of project goals and divergent perspectives
can impede the team’s ability to reach consensus on project goals and priorities. In
such instances, it is imperative that the variety of expertise in a team also be pertinent
to the project. This might be especially the case in ISD projects, where team pro-
cesses involve more complex forms of interdependence among team members rela-
tive to, say, industrial work teams. In summary, expertise heterogeneity is likely to
provide a team more opportunities for creatively combining a variety of knowledge
and perspectives, while simultaneously posing a challenge in the team’s ability to
reconcile excessively divergent ideas when heterogeneity is high. This leads to our
second hypothesis.
H2: Expertise heterogeneity is positively related to expertise integration in ISD
teams.
Relational Capital
Relational capital is defined as the level of trust, reciprocity, and closeness of work-
ing relationships among the members of a team [35]. Integrating a given team member’s
expertise into the team’s development activities requires that others in the team both
trust his or her expertise and be able to incorporate it with relative ease. Relational
capital facilitates this. Higher levels of relational capital increase the likelihood that
individuals in the team trust each other [64], which in turn, increases their willingness
01 tiwana.pmd 6/15/2005, 11:21 PM21
22 AMRIT TIWANA AND EPHRAIM R. MCLEAN
to build on each other’s perspectives, ideas, and expertise during the ISD process.
The close working relationship facet of relational capital reduces the costs of doing
so, given that stronger ties are associated with lower costs of sharing and eventually
integrating complex tacit knowledge [29]. Recall also that the precise expertise con-
tributions of each individual can be difficult to predict ex ante, although the general
nature of each individual’s contributions is predictable based on her assigned role in
the project. The reciprocity facet of relational capital facilitates contributions of ex-
pertise beyond levels that can be negotiated in advance [45]. Higher levels of rela-
tional capital will thus enhance team creativity by facilitating project-level integration
of diverse ideas, perspectives, and expertise that individual team members bring to
the project. This leads to our third hypothesis.
H3: Relational capital is positively and directly related to expertise integration
in ISD teams.
Absorptive Capacity
Absorptive capacity is defined as the ability of the members of a team to interrelate
with the expertise of their peer team members [13, 67]. Although the construct of
absorptive capacity has been studied mostly at the firm level, Cohen and Levinthal’s
[13] original description also explicitly identifies the construct at the group level.
Although a team’s absorptive capacity stems from the individuals belonging to it,
they caution that it is not resident in any single individual but depends on a mosaic of
individual capabilities. A team’s absorptive capacity in a given knowledge domain is
therefore a function of the expertise of the individuals in that domain and the exper-
tise of other team members with whom such individuals are collaborating.9 Absorp-
tive capacity at the team level raises individuals’ awareness of the pool of available
expertise in their team, lowers interpretive ambiguities, and breaks individual team
members from confinement in their own thought worlds, as discussed next.
First, integration of complex technical knowledge requires that team members be
competent in their individual areas of expertise but also familiar with the expertise
and skills of others in their team [13, 52]. Members of a team with higher absorptive
capacity are more likely to hold shared conceptualizations of each other’s expertise.
This allows team members to recognize how their peers’ expertise complements their
own, especially when it falls outside their own speciality.
Second, the same information is subject to multiple plausible interpretations. For
example, project requirements are derived from expressed or inferred customer needs
and business goals [3]. Interpretations of such information are socially constructed
by the members of a project team. But, using any such information requires that the
receivers possess sufficient background knowledge to understand it. The cognitive
limits of individual knowledge, however, force them to rely on their peer team mem-
bers to help interpret and integrate such information [27]. Individual team members
therefore serve as knowledge brokers or interpreters of information that relates to
their area of expertise, thereby making it more intelligible to the entire team. In a
01 tiwana.pmd 6/15/2005, 11:21 PM22
EXPERTISE INTEGRATION AND CREATIVITY IN INFORMATION SYSTEMS DEVELOPMENT 23
qualitative field study, Curtis et al. [16] found that individuals with specialized ex-
pertise often act as hubs for bringing outside expertise into the project team. In con-
trast, critical information held by one team member might never be utilized by a team
that lacks the capacity to interpret it. Therefore, higher levels of absorptive capacity
will assist the team in the interpretation of project-related information, raising its
prospects for integration during the ISD process. This leads to our fourth hypothesis.
H4: Absorptive capacity is positively related to expertise integration in ISD teams.
The Mediating Role of Expertise Integration
Finally, we posit that the effects of expertise heterogeneity, relational capital, and
absorptive capacity on creativity are mediated by expertise integration. Consider each
in turn. Diverse individually held expertise enhances the creativity of software project
teams primarily because such expertise can potentially lead to its project-specific
recombinations in less obvious ways. Although the positive association between the
diversity of viewpoints in a team and creativity has been known since the early 1960s,
such heterogeneity by itself does not lead to higher levels of creativity. For example,
Dougherty’s [21] study of cross-functional product development teams found that
teams perform better when their members combine their perspectives in an interac-
tive, iterative fashion. Similarly, relational capital enhances creativity because it al-
lows team members to experiment with new recombinations of tacit knowledge
embodied in individual expertise, viewpoints, and ideas. Likewise, absorptive capac-
ity enhances creativity because it helps find novel linkages between disparate ideas,
perspectives, and knowledge held by individual team members. For the foregoing
reasons, we hypothesize that these variables influence creativity primarily because
they facilitate expertise integration. This leads to our final hypotheses.
H5a: The influence of expertise heterogeneity on an ISD project team’s creativity
is fully mediated by expertise integration.
H5b: The influence of relational capital on an ISD project team’s creativity is
fully mediated by expertise integration.
H5c: The influence of absorptive capacity on an ISD project team’s creativity is
fully mediated by expertise integration.
Methodology
A FIELD SURVEY WAS CONDUCTED to test the hypothesized relationships. Figure 1
provides an overview of the methodology. Project-level data on each project were
collected from multiple stakeholders in each project team. The respondents included
multiple members from each project team as well as the primary user-side manager
from the organization that eventually used the system that was developed by each
project team. This approach was appropriate for the study, because the objective was
01 tiwana.pmd 6/15/2005, 11:21 PM23
24 AMRIT TIWANA AND EPHRAIM R. MCLEAN
to empirically test the proposed model, which itself was built on a synthesis of prior
work that had adopted an observational, qualitative approach for examining knowl-
edge management issues in ISD projects.
Description of the Research Setting and Data Collection
We gained access to a sample of interorganizational software projects through man-
agement sponsorship in a large U.S. services conglomerate. All teams in our study
were created temporarily for the project and were collocated for its entire duration.
The typical project lasted about six months, and some were as short as eight weeks.
The short duration of these projects afforded little time for knowledge transfer among
team members, as discussed earlier in the theory section. Therefore, the sample was
Figure 1. An Overview of the Study Methodology
01 tiwana.pmd 6/15/2005, 11:21 PM24
EXPERTISE INTEGRATION AND CREATIVITY IN INFORMATION SYSTEMS DEVELOPMENT 25
appropriate to examine the issue of expertise integration, since it was more likely to
be salient in such relatively shorter projects. We approached 173 individuals working
in 46 different project teams with the survey questionnaire.10 Data were collected
from multiple team members in each project team. We received 142 responses from
individuals in 42 project teams after two reminders. This represents an 82 percent
(142 responses/173 requests) response rate at the individual level and 91 percent (42
projects for which responses were obtained/46 projects whose members were con-
tacted) at the project level. Following this round of data collection, project success
evaluations for each project were obtained from a senior manager who spearheaded
the project.
Construct Operationalization and Scale Development
All research variables were measured using multi-item five-point Likert scales, as
summarized in Table 1. Since this study involves constructs that exist at multiple
levels of analysis, we paid close attention to ensuring that all of our scales were
operationalized at the team level (see [37]).
Expertise heterogeneity was measured using Campion et al.’s [10] three-item scale
that assesses the extent to which team members represent diverse backgrounds, expe-
riences, and nonoverlapping domains of expertise. Relational capital was measured
using a team-level adaptation of Kale et al.’s [35] five-item scale. This scale assesses
member’s perceptions of team-level trust, reciprocity, and strength of ties. Creativity
was assessed by tapping into the extent to which the team’s processes were creative.
An outcomes-oriented approach in a variance model would instead have measured
the extent to which the team’s outcomes were judged as being creative. We used
Table 1. Summary of Key Constructs and Their Measures
Number of Source of
Construct Definition items scale
Creativity The degree to which a project team’s 3 [19]
processes are novel in the context of
the project’s objectives.
Expertise Synthesis of individually held specialist 4 New scale
integration expertise at the project level.
Expertise The diversity in the expertise possessed 3 [10]
heterogeneity by the members of a project team.
Relational The level of level trust, reciprocity, and 5 [35]
capital proximity of ties among the members of
a project team.
Absorptive The ability of the members of a team 3 New scale
capacity to interrelate in a project’s context to the
expertise of their peers outside of their
own specialized domain.
01 tiwana.pmd 6/15/2005, 11:21 PM25
26 AMRIT TIWANA AND EPHRAIM R. MCLEAN
Denison et al.’s [19] three-item measure that assessed the team’s experimentation
with alternative ways of solving the problem at hand, the extent to which a team was
imaginative in thinking about new and better ways, and invention of new ways to
address nonroutine project matters.11
Scales to measure expertise integration and absorptive capacity were developed
through a multistep, iterative procedure.12 We measured expertise integration by as-
sessing the extent to which a team’s members synthesized their individual expertise
at the project level, synthesized various members’ tacit knowledge and expertise in
developing project concepts, understood the project from a systemic perspective, and
synthesized their own expertise with such project-level knowledge [27, 52, 70].13 We
measured absorptive capacity as a team-level construct that assessed the extent to
which individual members in each project team could interrelate with the knowledge,
expertise, and skills of their peer team members (e.g., [8, 13, 75]).
We controlled for two variables that might influence team creativity: (1) techno-
logical uncertainty and (2) stage of the project at which data were collected from
team members. Technological uncertainty is defined as the rapidity with which the
hardware and software central to the project and the skill sets associated with them
were changing. A project context characterized by high technological uncertainty is
more likely to evoke creative responses to project problems. The scale was adapted
from Poppo and Zenger’s [54] technological uncertainty scale that was developed in
an IS context. As for project stage, as a project approaches its very last stages, a team
is less likely to engage in creative processes and will instead focus on tying loose
ends before looming deadlines approach. The measure asked the project manager
(not the project sponsor who evaluated the project’s success) to assess the current
stage of the project on a five-point scale with 20 percent increments.14
Limitations
Before discussing the results, five limitations of this study should be noted. First, the
model did not cover all antecedents of team creativity. Instead, it focused only on a
subset of factors that influence expertise integration. Second, the correlations be-
tween relational capital and expertise integration were high. Although measurement
model analyses showed that there is sufficient discriminant validity between these
constructs, caution must be exercised in interpreting the associated results. Third,
project success assessments from senior managers were used as a secondary check
for common methods bias. Since this variable is downstream from creativity, it only
partially addresses this concern. Nevertheless, since the main dependent variable—
team creativity—was based on assessments of multiple team members in each project,
common methods bias is not likely to be a persuasive threat. Fourth, lack of support
for the hypothesized positive relationship between expertise heterogeneity and ex-
pertise integration warrants further attention. It is difficult to rule out the possibility
that the absence of the significant relationship might be an artifact of the way in
which the construct was measured. Finally, creativity was measured as the degree to
which a team’s processes were creative, consistent with the variance modeling ap-
01 tiwana.pmd 6/15/2005, 11:21 PM26
EXPERTISE INTEGRATION AND CREATIVITY IN INFORMATION SYSTEMS DEVELOPMENT 27
proach used in this study. This however is not the same as truly process observations,
which require longitudinal data.
Results
WE USED PARTIAL LEAST SQUARES (PLS)—a second-generation structural equation
modeling technique—to revalidate the measurement model and then to test the hy-
pothesized relationships in two steps using PLS-Graph 2.91.15
Measurement Model Assessment
Convergent and discriminant validity for all scales was assessed both before and after
aggregating data to the team level (only postaggregation values are reported for brev-
ity). The correlations among various constructs shown in Table 2 are the team-level,
postaggregation values. As Table 2 shows, each item loaded highly on its respective
construct. Cronbach’s alpha for all constructs exceeds the 0.7 threshold recommended
by Nunally [48], and the internal consistency reliability (ICR) for each construct
exceeds 0.8, confirming convergent validity [25]. The loadings of all indicators on
the corresponding theoretical constructs exceed the recommended 0.7 threshold in
the PLS measurement model (see Table 2) [12, 33]. Cronbach’s alpha is not com-
puted for technological uncertainty because it has two measurement items. Discrimi-
nant validity was indicated by three assessments in the PLS measurement model: (1)
items had low (< 0.5) and nonsignificant cross-loadings, (2) the diagonal elements
representing √ρvc exceeded the off-diagonal elements in the construct correlation
matrix, and (3) the ratio of the variance in the indicators for each construct relative to
the total amount of variance ρvc exceeded 0.5 [25]. For example, the bold diagonal
element for expertise integration is 0.92, which exceeds all off-diagonal elements in
Table 2. Similarly, relational capital has a diagonal element of 0.86, which again
exceeds all off-diagonal elements in the correlation matrix. This assessment along
with ρvc values of 0.85 and 0.74, respectively, (both of which exceed 0.5) suggests
that in spite of the correlations observed in the exploratory factor analysis (shown in
the Appendix), the constructs exhibit sufficient discriminant validity. Similar pat-
terns of results for expertise heterogeneity and absorptive capacity suggest that the
constructs exhibit sufficient discriminant validity.
Data Aggregation
Data were collected from multiple respondents for each project team and were aggre-
gated for each project after assessing within-team agreement. To assess such within-
team agreement, the interclass correlation coefficient (ICC) was used to test whether
membership in a team also led to similar patterns of responses. The ICC values re-
ported in Table 1 range from 0.58 to 0.87, which indicates sufficient within-team
agreement to justify aggregation [36]. Note that all constructs are operationalized at
01 tiwana.pmd 6/15/2005, 11:21 PM27
28 AMRIT TIWANA AND EPHRAIM R. MCLEAN
Table 2. Psychometric Properties and Aggregation Statistics of Key Constructs
Number of
items (PLS
Mean S.D. loadingsa)αiαtICR ICC ρvc 12345678
1. Expertise 3.77 0.54 4 (0.87, 0.92 0.94 0.96 0.74 0.85 0.92
integration 0.94, 0.95,
0.93)
2. Creativity 3.63 0.69 3 (0.92, 0.80 0.86 0.91 0.58 0.71 0.71 0.84
0.80, 0.93)
3. Exper tise 3.75 0.43 3 (0.98, 0.95 0.92 0.95 0.87 0.91 –0.19 –0.18 0.95
heterogeneity 0.89, 0.93)
4. Absorptive 3.91 0.67 3 (0.94, 0.93 0.93 0.95 0.82 0.55 0.76 0.6 –0.18 0.74
capacity 0.97, 0.88)
5. Relational 3.74 0.81 5 (0.86, 0.91 0.93 0.96 0.73 0.74 0.70 0.53 –0.04 0.55 0.86
capital 0.91, 0.94,
0.92, 0.92)
6. Project 3.71 0.37 3 (0.62*, — 0.8 0.83 — 0.63 0.18 0.21 0.18 0.19 0.24 0.79
successb0.91, 0.82)
7. Technological 2.94 0.61 2 (0.98, — — 0.94 0.86 0.81 0.11 0.13 –0.27 0.18 0.1 –0.23 0.9
uncertainty 0.90)
8. Project 80–100 1.42 Single item —————0.17 –0.01 0.27 0.25 0.15 0.35 –0.07 —
stage percentcmeasure
Notes: Shaded diagonal elements are the square root of the shared variance between the constructs and their measures. Off-diagonal elements are the correlations between the different constructs
(aggregated team-level data). ICR = Fornell and Larcker’s (1981) internal consistency reliability. a p < 0.001; * p < 0.01. b Senior manager’s evaluation. c Median value on a five-point scale with 20
percent increments. αi and αt represent pre- and postaggregation construct reliability coefficients.
01 tiwana.pmd 6/15/2005, 11:21 PM28
EXPERTISE INTEGRATION AND CREATIVITY IN INFORMATION SYSTEMS DEVELOPMENT 29
the team level; aggregated values are therefore the average of individual members’
perceptions of their team. Since only one response was available for each project’s
success, no such aggregation was possible for this construct.
Descriptive Statistics and Demographics
The respondents in our sample had about 7.8 years of information technology (IT)
experience and had been with their organization for about 16 months on average. The
average team consisted of about nine members. The average duration of the projects
was 6.2 months. At the time data were collected from project team members, most
projects were in their final stages (the median value of the project stage variable was
“80–100 percent complete”).
Structural Model Assessment
A PLS structural model represents the relationships among various latent constructs.
Paths in this model are interpreted as standardized regression weights and the load-
ings on each construct as loadings in principal component analyses. A bootstrapping
procedure with replacement using 500 subsamples was used to estimate the statistical
significance of the parameter estimates. A summary of these results is presented in
Figure 2.
Expertise integration had a significant positive effect on creativity (β = 0.719,
t-value = 7.616, p < 0.001), supporting H1. H2, which posited a positive relationship
between expertise heterogeneity and expertise integration, was not supported (β=
–0.059, t-value = –1.08, p > 0.1). Relational capital had a significant positive effect
on expertise integration (β = 0.349, t-value = 3.76, p < 0.001), supporting H3. Fi-
nally, absorptive capacity had a significant positive effect on expertise integration
(β= 0.645, t-value = 7.39, p < 0.001), supporting H4. To test the mediation role of
expertise integration, we assessed the direct effects of expertise heterogeneity, rela-
tional capital, and absorptive capacity on creativity. The effects of expertise hetero-
geneity (β = –0.023, t-value = –0.228, p > 0.1), relational capital (β = 0.100, t-value =
0.728, p > 0.1), and absorptive capacity (β = 0.355, t-value = 1.365, p > 0.1) on
creativity all lacked statistical significance. Thus, the full mediation hypotheses, H5b
and H5c, were supported. Absence of any significant direct relationships between
these variables and creativity confirms full mediation by expertise integration and
lends support to our argument that they influence creativity primarily because they
facilitate expertise integration. As expected, technological uncertainty had a signifi-
cant positive relationship with creativity (β = 0.246, t-value = 2.154, p < 0.05). The
other control variable, project stage, was not significant (β = –0.14, t-value = –1.492,
p > 0.10). Although it is difficult to establish causality using cross-sectional data, the
absence of any significant direct effects of expertise heterogeneity, relational capital,
and absorptive capacity on creativity provide some support for a causal association.
We also assessed the threat of common methods bias and confirmed that creative
01 tiwana.pmd 6/15/2005, 11:21 PM29
30 AMRIT TIWANA AND EPHRAIM R. MCLEAN
processes at the team level do lead to superior project outcomes.16 Table 3 summa-
rizes these results.
Model Quality Assessment
The predictive quality of a model can be assessed in two ways: (1) percentage of the
total variance it explains (R2) and (2) its predictive relevance score (Q2). Our model
explained 58.8 percent of the variance in creativity (R2 = 0.588). Of this, 7.1 percent
of the variance was explained by the two control variables and the remaining 51.7
percent was explained by expertise integration. In addition, the model explained 80.7
percent of the variance in expertise integration, suggesting that relational capital and
absorptive capacity are powerful predictors of expertise integration. Furthermore, the
average Q2 value across all five runs was 0.3553, which suggests that our model has
high predictive relevance [11].17 Together, the R2 and Q2 values suggest that the model
predicts creativity reasonably well.
Discussion and Implications
THIS STUDY WAS MOTIVATED BY the scantness of prior research on team-level cre-
ativity in the ISD process, which is becoming ever more critical as organizations seek
to develop software applications in new and novel problem domains. A noteworthy
feature of our approach is that we conceptualized creativity in the ISD process in-
stead of simply judging whether the outcomes of the project were creative. This re-
Figure 2. Research Model and Results
Notes: ----- Nonsignificant path; significant paths are in boldface; *** p < 0.001,
** p < 0.01, * p < 0.05.
01 tiwana.pmd 6/15/2005, 11:21 PM30
EXPERTISE INTEGRATION AND CREATIVITY IN INFORMATION SYSTEMS DEVELOPMENT 31
fined conceptualization is consistent with the emerging contemporary view of cre-
ativity in management and psychology.
This is one of the first studies to develop a conceptual explanation of the mecha-
nism of knowledge integration through which individually held expertise in a team
translates into creativity at the team level. We further developed an explanation for
the antecedents of creativity, which we argued influence creativity by facilitating
integration of individually held expertise at the project level. Our conceptualization
of ISD project teams as systems of relationships and expertise, use of teams as the
unit of analysis, a field setting, and use of multiple informants for each project are
noteworthy. The results (see Table 3) provide considerable support for the proposed
idea that expertise integration is the key explanatory mechanism through which indi-
vidually held expertise leads to creativity at the team level.
The Relationship Between Expertise Integration and
Team Creativity
The first important finding in our analysis is the positive and significant relationship
between expertise integration and creativity (H1). The results suggest that team cre-
ativity results from developing novel associations and linkages among the diverse
ideas, perspectives, and domain expertise that individual team members bring to the
project. A large path coefficient of about 0.72 between expertise integration and cre-
ativity implies that teams that can draw on their members’ expertise in ways that
Table 3. Summary of Hypothesis Tests
Hypothesized β
Hypothesis effect Supported (T-value)
H1: Expertise integration → Creativity + Yes 0.72***
(7.62)
H2: Expertise heterogeneity → Expertise + No –0.06
integration (–1.1)
H3: Relational capital → Expertise + Yes 0.35**
integration (3.76)
H4: Absorptive capacity → Expertise + Yes 0.64***
integration (7.39)
Full mediation hypothesis
H5a: Expertise heterogeneity → Creativity Full mediation No –0.23
by expertise (–0.29)
integration
H5b: Relational capital → Creativity Yes 0.10
(0.73)
H5c: Absorptive capacity → Creativity Yes 0.35
(1.36)
* p < 0.05, ** p < 0.01, *** p < 0.001.
01 tiwana.pmd 6/15/2005, 11:21 PM31
32 AMRIT TIWANA AND EPHRAIM R. MCLEAN
allow individuals to build on each other’s knowledge, skills, and perspectives are
more likely to be creative. Expertise integration alone explained over 50 percent of
the variance in creativity, suggesting that integration of individually held expertise in
a team is an important predictor of an ISD team’s creativity. Such integration allows
the diverse perspectives and skills within a team to be brought together in a fashion
that directly contributes to the creativity of the team in the ISD process. This finding
is the first to test the idea that higher levels of creativity are the result of different
cognitive structures and skill sets coming together. This result has profound implica-
tions for how organizations must approach the organizing of teams for creativity-
seeking ISD projects. Clearly, bringing together the right set of skills is insufficient
for enhancing creativity in the ISD process; it also requires that those skill sets be
appropriately integrated and brought to bear on the development process. We also
demonstrated empirically that team creativity is positively associated with superior
project outcomes.18
The Mediating Role of Expertise Integration
The second important finding is that the expertise integration plays an important
mediating role between the presence of expertise in a project team and its creativity in
the ISD process. Our perspective details a team process whereby synergy across dis-
parate knowledge bases can lead to more creative ISD processes.
We had predicted that the presence of diverse expertise in a team, members’ ability
to interrelate with each other’s expertise, and good working relationships within the
team are important for creativity because they facilitate integration of their expertise
at the project level. The absence of any direct effects of these compositional charac-
teristics of team expertise and relational characteristics (H5a, Hb, and Hc) supports
this logic. This finding shows that the compositional characteristics of an ISD team’s
expertise influence its creativity primarily through the process of expertise integra-
tion. Prior research on knowledge management in ISD projects has never before es-
tablished this linkage, although the concept of integration has been alluded to as
being important [24, 70]. This is a novel insight, because it identifies a key mecha-
nism through which expertise in a team translates into creativity during the ISD pro-
cess. This finding cautions managers that merely collecting a variety of relevant
expertise in a project team is not sufficient for creativity to emerge; managers can
collect all the relevant skills, expertise, and experience bases in a team yet fail to
stimulate creativity unless they conduce project-specific integration of team mem-
bers’ expertise.
The concept of expertise integration and the empirical support that it received in
our analyses brings into question an implicit assumption that problem domain knowl-
edge can be transferred to the ISD team during the up-front requirements gathering
phase of the project. This has often resulted in requirements that either end up being
unstable or incomplete, exposing the project to the risk of misfit between the users’
needs and the delivered system. Our results highlight the need to consider expertise
integration as a viable mechanism instead of relying too heavily on the assumption
01 tiwana.pmd 6/15/2005, 11:21 PM32
EXPERTISE INTEGRATION AND CREATIVITY IN INFORMATION SYSTEMS DEVELOPMENT 33
that knowledge about customer needs can easily and accurately be transferred. Nota-
bly, no prior study that has explored the idea of integration in ISD projects has devel-
oped such a measure for the concept. Our study developed and validated a new
team-level measure for expertise integration, which might be readily usable in ex-
ploring the concept in future research.
Compositional Antecedents of Expertise Integration and
Creativity
The third important set of findings in this study relate to how the compositional at-
tributes of an ISD project team’s design influence its creativity through the process of
expertise integration. Since this influence was fully mediated by expertise integra-
tion, we restrict the scope of this discussion only to their influence on expertise inte-
gration. Including the antecedents of expertise integration has theoretical importance
because it provides some new insights into how teams that seek creativity ought to be
constituted. We focused on a subset of three such attributes—expertise heterogene-
ity, relational capital, and absorptive capacity—that most directly influence the abil-
ity of individuals to utilize their peer team members’ expertise for project tasks and
the relationships through which expertise integration is facilitated. (These correspond
with H2, H3, and H4.) Our findings not only lend support to, but also extend,
Szulanski’s [64] guiding logic for the choice of these variables.
Relational Capital
The observed positive and significant relationship between relational capital and ex-
pertise integration suggests that accessibility of other individuals’ expertise within a
team is an important predictor of its application to the project, especially when a
detailed breakdown of each member’s contributions cannot be fully anticipated in
advance. Since relational capital might accrue from interactions over time, it is pos-
sible that teams with history will have higher levels of relational capital. From a
theoretical perspective, future research should also examine whether there are any
downsides to having too much relational capital. A potential theoretical lens could be
the strength of ties literature, which suggests that the same strong ties that ease knowl-
edge sharing might also reduce the prospect for novel recombinations. From a prag-
matic perspective, this finding suggests that managerial interventions that build
relational capital in newly formed teams can enhance creativity.
Absorptive Capacity
The positive influence of absorptive capacity on expertise integration points to the
value of individual team members’ having a basic understanding of the knowledge
domains of other individuals with when they interact over the course of a project.
Such understanding enhances their awareness of where complementary expertise re-
sides within their team, reduces the likelihood of misinterpretation of project-related
01 tiwana.pmd 6/15/2005, 11:21 PM33
34 AMRIT TIWANA AND EPHRAIM R. MCLEAN
information outside of their immediate domain, and allows the team to experiment
with new combinations of ideas, perspectives, and concepts.19 This finding cautions
managers against perfect partitioning of project tasks among individual team mem-
bers under the presumption that it improves development efficiency, because it
unintendedly might stifle creativity. A larger path coefficient between absorptive ca-
pacity and expertise integration compared to that between relational capital and ex-
pertise integration indicates that absorptive capacity has a more pronounced effect on
a team’s capacity for integrating its members’ expertise. However, the absolute mag-
nitude of the path from relational capital is fairly large at about 0.35, indicating that
relational capital is critical to the creativity of ISD project teams.
Expertise Heterogeneity
The lack of support for the predicted positive relationship between expertise hetero-
geneity and expertise integration is a surprising result. The negative direction of this
path coefficient warrants further discussion. Although much of the contributing lit-
erature led us to predict a heterogeneous collection of expertise in a team as being
unconditionally helpful, these results indicate that that might not always be the case.
There are two possible interpretations for this result: (1) high levels of dissimilarity
among individually held expertise might raise the arduousness of resolving conflict-
ing ideas held by individual members to a point where the benefits of access to di-
verse expertise outweigh the costs of using it, or (2) previous theory is not completely
applicable in this context. Both these interpretations are discussed next.
First, individuals in highly heterogeneous teams have fewer overlaps in their knowl-
edge.20 In the absence of sufficient overlaps, they might have difficulty in discussing
the knowledge that they uniquely hold [56]. Although heterogeneity brings a more
varied pool of expertise for potential recombination, the difficulties in reconciling
diverse interpretations of project goals and different perspectives on possible solu-
tions might impede the team’s ability to reach consensus on project goals and priori-
ties. This interpretation is consistent with Cooper’s [15] suggestion that increased
diversity in ISD groups can decrease goal congruence.
Second, the existing body of research that suggests a positive relationship between
heterogeneity and team processes was developed largely in industrial teams such as
those found in assembly lines and manufacturing plants (see, e.g., [19, 20, 40]). The
tasks of such teams are less knowledge intensive, involve lower levels of expertise
interdependence (i.e., they rely on simple pooling or sequential application of indi-
vidual expertise), and their outputs are less contingent on integrating the knowledge
of various members. Creativity theories remain largely untested in uncertain and rap-
idly changing environments [60]. Therefore, the theory behind the hypothesized posi-
tive relationship might not be generalizable to ISD teams that typically are engaged
in knowledge-intensive tasks that involve more complex forms of interdependence.
Overall, the pattern of results suggests that the composition of a team is a critical
factor in determining whether that team is likely to be creative in the ISD process.
Managers must clearly consider both the relationships among team members and the
01 tiwana.pmd 6/15/2005, 11:21 PM34
EXPERTISE INTEGRATION AND CREATIVITY IN INFORMATION SYSTEMS DEVELOPMENT 35
knowledge complementarities that align their work at the project level. Although a
variety of organizational constraints might limit the discretion that managers can ex-
ercise in manipulating team design from these perspectives, some sensitivity to these
issues can inform team constitution at the outset of an ISD project.
Key Contributions
The study makes three novel contributions to the IS literature. We first developed and
tested the idea that individually held expertise influences creativity in the ISD pro-
cess primarily through the process of expertise integration at the team level—a rela-
tionship that has not previously been examined. Our finding of full mediation of the
limited set of antecedents of creativity by expertise integration supports this idea.
This contribution integrates two disparate bodies of literature on knowledge in the
ISD process and the understudied issue of creativity. The second contribution of this
paper is in showing that expertise integration enhances creativity in the ISD process.
With the exception of one qualitative study [15], no prior study has even raised the
role of team-level expertise integration in enhancing ISD team creativity. A large
proportion of variance in creativity explained by expertise integration points to its
centrality in explaining team creativity. We also built on prior descriptive studies on
the integration of knowledge in ISD projects [16, 70] to develop and validate a new
team-level measure for the construct of expertise integration. The third contribution
of the study is showing that the compositional and relational attributes of a team’s
expertise influence the extent to which its members can integrate their diverse exper-
tise bases in formulating a coherent project solution, which in turn influences creativ-
ity at the project level. Such compositional attributes have seldom been considered in
prior studies of project teams [56]. These attributes were the level of heterogeneity of
team members’ expertise, the quality of the working relationships within the team,
and the team’s absorptive capacity. The logic developed and validated was that these
variables predict the degree to which the disparate expertise held by the individuals in
a team can be integrated at the project level. In summary, relational capital and ab-
sorptive capacity do not influence creativity in software project teams in and of them-
selves; they do so because they facilitate expertise integration. This understanding of
the factors that shape creative processes in teams and our findings related to the me-
diating role of expertise integration provides managers some basic tools to enhance
creativity both by better structuring teams and by deploying collaborative develop-
ment tools that support expertise integration during systems development.
Conclusions
THE OVERARCHING GOAL IN THIS PAPER was to examine how the expertise of indi-
vidual team members translates into creativity in the ISD process. We theoretically
developed the idea that the key process through which this happens during the ISD
process is via the integration of individually held expertise at the project level. We
also identified three attributes of the team’s composition that were likely to influence
01 tiwana.pmd 6/15/2005, 11:21 PM35
36 AMRIT TIWANA AND EPHRAIM R. MCLEAN
expertise integration at the team level, and in turn, the team’s creativity during the
ISD process. We tested the proposed model using data collected from a field study of
142 individuals in 42 software projects. ISD project teams as the unit of analysis and
the use of multiple respondents for each project are noteworthy strengths of the study.
Our team-level conceptualization especially complements prior research that has fo-
cused largely on personal factors that influence individual creativity.
The paper contributes three new insights into creativity in ISD project teams. First,
it theoretically links expertise integration with team creativity. Second, it shows that
the ability of a team’s members to interrelate with the expertise of others in the team
and good working relationships are critical for creativity. The surprising lack of rela-
tionship between heterogeneity of expertise and team creativity raises issues for fu-
ture inquiry. While creativity is often portrayed as something that cannot be defined,
implemented, or created, our findings suggest that IS managers can indeed design
ISD teams to be more creative, albeit within organizational constraints. Third, it shows
that expertise integration is the key mechanism through which individual expertise
results in team creativity during the ISD process. Although this study provides a
starting point for research into the domain of team-level creativity in the ISD process,
more exhaustive team-level models must be developed before we can fully grasp
creativity in ISD.
Acknowledgments: The authors gratefully acknowledge input from Ashley Bush, Benn
Konsynski, Mark Keil, Anandhi Bharadwaj, Bala Ramesh, Arun Rai, and three anonymous
reviewers.
NOTES
1. We are grateful to an anonymous reviewer for suggesting this point.
2. The prevalent tradition of creativity research in the reference disciplines such as man-
agement and psychology has largely been at the individual level, focusing primarily on indi-
vidual differences and antecedent conditions for individual creativity [74]. But teams rather
than individuals usually develop software systems. Since ISD involves the collective creative
processes of many team members, it is necessary to adopt a group-level perspective on creativ-
ity [15].
3. In this study, we measure some process-related attributes of creativity, since the model
itself is a variance model; studying the process itself requires a process model. Therefore, we
assess the degree to which the team’s processes were creative in our measurement of the con-
struct.
4. Although Grant’s [27] conceptualization of expertise integration is anchored at the firm
level, Grant explicitly recognizes that organizations merely provide the context for integrating
expertise and that the actual integration is carried out in teams composed of individuals. We
use the term expertise to acknowledge the presence of tacit as well as explicit knowledge.
Notably, it is individually held tacit knowledge that is most closely associated with ISD tasks
[55]. Such knowledge involves an understanding of the application problem domain as well as
technical knowledge.
5. Explicit knowledge is also referred to as declarative knowledge in the software develop-
ment literature. Declarative knowledge is defined as knowledge that is static and based on facts
concerned with objects, persons, and events and their relationships [1, 55]. Tacit knowledge
involved in the software development process is similarly referred to as contextual, application
domain, and procedural knowledge [55, 57]. Robillard’s description of such knowledge—it is
01 tiwana.pmd 6/15/2005, 11:21 PM36
EXPERTISE INTEGRATION AND CREATIVITY IN INFORMATION SYSTEMS DEVELOPMENT 37
difficult to describe, but once learned, is rarely forgotten—mirrors Nonaka’s definition of tacit
knowledge.
6. Such need for integrating knowledge about the problem domain is reflected in the par-
ticipatory design philosophy for ISD, which emphasizes the active involvement of potential
users of the system in the design and development process of the system.
7. Although Szulanski [64] did not explicitly identify these variables, he suggested that
poor integration is caused by the inability to build on other individuals’ knowledge and by the
poor quality of relationships among them that make integration difficult. By corollary, teams
with compositional attributes that are conducive to integrating individually held expertise at the
project level are more likely to be effective in achieving higher levels of expertise integration.
8. For example, a team might be able to validate its interpretation of customer requirements
through its access to other external groups that might have previously worked with that cus-
tomer. Or, a team might get feedback on the choice of a particular programming language or
design approach that is new to the team but has been used elsewhere in one team member’s
extended network.
9. Expertise heterogeneity and absorptive capacity are conceptually distinct constructs:
although a team might have heterogeneity in its members’ expertise, it does not necessarily
follow that the members of the team can relate with each other’s expertise. A simple example
is a two-person team composed of an accountant and a programmer. Although they have high-
expertise heterogeneity, it does not automatically imply that the accountant can relate with
programming or that the programmer can understand accounting (i.e., each other’s domains).
Therefore, a team can have high expertise heterogeneity yet be low on absorptive capacity.
Similarly, it is also possible that the same team above has a programmer who has a basic
understanding of what the accountant does and vice versa (i.e., absorptive capacity is high).
10. Data were collected using a Web-based questionnaire. A list of 46 active projects was
coded into the questionnaire. Each project had a unique identification number known to all
respondents; this was also included in the project list. We electronically distributed the ques-
tionnaire by sending 173 individuals e-mail messages with a personalized URL for the survey.
11. In their original study, this measure assessed creative strategy that was a team-level
process-focused measure. Although it bears some similarities to earlier individual-level mea-
sures of creativity, there is considerable theoretical support for the similarities in the concep-
tual structure of creativity at individual and group levels [22, 74].
12. We identified potential items for both scales beginning with a comprehensive review of
the literature. We refined this initial item pool through multiple rounds of feedback from three
managers and six academic domain experts. We conducted a review session with these experts
before the first pretest, and then again after the first and second pretest phases. We adminis-
tered the instrument to a convenience sample of 79 semester-long software project teams in
three sequential sessions involving 25, 21, and 33 teams. Only a small subset of the potential
scale items from the initial pool were retained. Various items were refined and simplified
based on their feedback and the results of exploratory factor analyses. The final scales for
absorptive capacity and expertise integration consisted of three and four items, respectively,
and were revalidated with the main data set.
13. Our use of Grant’s [27] conceptualization of knowledge integration is somewhat limited
in the sense that our measure does not tap into the various mechanisms by which knowledge
can be integrated. Instead, the items in our scale measure it largely as an outcome, consistent
with its operationalization as a reflective construct (i.e., the scale items are caused by the latent
construct). A mechanisms-oriented scale would require a formative scale wherein the scale
items lead to knowledge integration. See Chin [11] for an exposition on the differences be-
tween reflective and formative measures.
14. It is, however, possible that the effect of project stage is not simply linear. Since our
objective was simply to control for the effects the stage of each project in the study, we used a
coarse, single-item measure to control for project stage. To prevent confounding effects from
a history of having worked together [65], we sampled only teams that were formed anew and
whose members did not have prior history of working together.
15. Our choice of PLS was guided by three considerations. First, two constructs in this study
use newly developed scales. PLS allows triangulation of the convergent and discriminant va-
lidity of these constructs with the traditional scale validation procedures used in the pretests.
01 tiwana.pmd 6/15/2005, 11:21 PM37
38 AMRIT TIWANA AND EPHRAIM R. MCLEAN
PLS’s ability to assess the measurement model within the context of its theoretical-mediated
model therefore makes it superior to multiple regression and path-analytic techniques. Second,
PLS is well suited for analyzing the smaller data set that resulted from aggregating individual
responses to the team level. Third, unlike LISREL, PLS makes no a priori normality assump-
tions regarding the data.
16. Since team creativity was assessed by multiple respondents for each project and suffi-
cient interrater agreement was observed among all respondents within each team, common-
methods bias is not a persuasive threat to the study. As a secondary check, correlations between
team creativity and project success assessments obtained for each project from the key external
project sponsor (who was a manager external to each project team) provided further assurance
that common-methods bias was not a persuasive problem. We used evaluations collected from
the upper-management sponsor for each project, each identified with the help of the three
sponsoring top managers and their choice corroborated with the key senior manager stake-
holder listed in the internal project roster. Following project success measures used in prior
studies of ISD [28, 31], these managers were asked to evaluate the extent to which the project
provided the desirable features and functionality, met its business objectives, and its overall
success at the time the completed project was delivered. Creativity had a significant positive
effect on project success (β = 0.338, t-value = 1.844, p < 0.05). While this relationship is not
central to this study, it provides empirical support that creative processes do contribute to
superior project outcomes. Although this does not entirely eliminate the threat of common-
methods bias, this result suggests that it is not a persuasive threat to our findings. Since project
success was assessed by a respondent different from the team’s members, and the data related
to the independent variables were collected from individual team members, this multisource
approach somewhat further mitigates the threat of common methods bias. Notably, this test
also empirically confirmed that creative team processes do contribute to superior project out-
comes.
17. We used a blindfolding procedure that omits part of the data for a given block of indica-
tors and then attempts to estimate the omitted part based on existing estimates [26]. Only prime
numbers less than the sample size can be used for such omission distances. We blindfolded
creativity using omission distances of 11, 13, 19, 31, and 41, and reestimated the model in five
separate runs. The Q2 estimates obtained were 0.3534, 0.3639, 0.3600, 0.3562, and 0.3430.
18. Recall that the projects in this study were novel Internet software applications. The
relationship between creativity and project success might be weaker in software maintenance
(as opposed to applications development) projects.
19. High levels of expertise heterogeneity might lower the absorptive capacity of a team
because heterogeneity reduces the expertise overlaps among individual team members. This
relationship is not formally hypothesized in this study. Nevertheless, in testing this relation-
ship, we found that the path coefficient was negative but the relationship lacked statistical
significance (β = –0.181, t-value = –1.56, p > 0.1).
20. The mean construct score of 3.75 on a five-point scale for expertise heterogeneity sup-
ports this assertion.
REFERENCES
1. Adelson, B., and Soloway, E. The role of domain experience in software design. IEEE
Transactions on Software Engineering, 11, 11 (1985), 1351–1360.
2. Alavi, M., and Tiwana, A. Knowledge integration in virtual teams: The potential role of
knowledge management systems. Journal of the American Society for Information Science
and Technology, 53, 12 (2002), 1029–1037.
3. Alexander, C. Notes on the Synthesis of Form. Cambridge: Harvard University Press,
1964.
4. Amabile, T.; Conti, R.; Coon, H.; Lazenby, J.; and Herron, M. Assessing the work
environment for creativity. Academy of Management Journal, 39, 5 (1996), 1154–1184.
5. Ancona, D.G., and Caldwell, D.F. Demography and design—Predictors of new product
team performance. Organization Science, 3, 3 (1992), 321–341.
01 tiwana.pmd 6/15/2005, 11:21 PM38
EXPERTISE INTEGRATION AND CREATIVITY IN INFORMATION SYSTEMS DEVELOPMENT 39
6. Barki, H.; Rivard, S.; and Talbot, J. An integrative contingency model of software project
risk management. Journal of Management Information Systems, 17, 4 (Spring 2001), 37–69.
7. Benaroch, M. Managing information technology risk: A real options perspective. Jour-
nal of Management Information Systems, 19, 2 (Fall 2002), 43–84.
8. Boynton, A.; Zmud, R.; and Jacobs, G. The influence of IT management practice on IT
use in large organizations. MIS Quarterly, 18, 3 (1994), 299–318.
9. Browne, G., and Rogich, M. An empirical investigation of user requirements elicitation:
Comparing the effectiveness of prompting techniques. Journal of Management Information
Systems, 17, 4 (Spring 2001), 223–250.
10. Campion, C.; Medsker, G.; and Higgs, A. Relations between work group characteristics
and effectiveness: Implications for designing effective work groups. Personnel Psychology,
46, 3 (1993), 823–855.
11. Chin, W. The partial least squares approach to structural equation modeling. In G.
Marcoulides (ed.), Modern Methods for Business Research. Mahwah, NJ: Lawrence Erlbaum,
1998, pp. 295–336.
12. Chin, W., and Newsted, P. Structural equation modeling analysis with small samples
using partial least squares. In R. Hoyle (ed.), Statistical Strategies for Small Sample Research.
Thousand Oaks, CA: Sage, 1999, pp. 307–341.
13. Cohen, W., and Levinthal, D. Absorptive capacity: A new perspective on learning and
innovation. Administrative Science Quarterly, 35, 1 (1990), 128–152.
14. Constantine, L., and Lockwood, L. Orchestrating project organization and management.
Communications of the ACM, 36, 10, (1993), 31–43.
15. Cooper, R. Information technology development creativity: A case study of attempted
radical change. MIS Quarterly, 24, 2 (2000), 245–276.
16. Curtis, B.; Krasner, H.; and Iscoe, N. A field study of the software design process for
large systems. Communications of the ACM, 31, 11 (1988), 1268–1287.
17. Darr, E.; Argote, L.; and Epple, D. The acquisition, transfer, and depreciation of knowl-
edge in service organizations: Productivity in franchises. Management Science, 41, 11 (1995),
1750–1762.
18. Dean, D.; Lee, J.; Pendergast, M.; Hickey, A.; and Nunamaker, J. Enabling the effective
involvement of multiple users: Methods and tools for collaborative software engineering. Journal
of Management Information Systems, 14, 3 (Winter 1996–97), 179–222.
19. Denison, D.; Hart, S.; and Kahn, J. From chimneys to cross-functional teams: Develop-
ing and validating a diagnostic model. Academy of Management Journal, 39, 4 (1996),
1005–1023.
20. Dooley, R., and Fryxell, G. Attaining decision quality and commitment from dissent:
The moderating effects of loyalty and competence in strategic decision-making teams. Acad-
emy of Management Journal, 42, 4 (1999), 389–402.
21. Dougherty, D. Interpretive barriers to successful product innovation in large firms. Or-
ganization Science, 3, 2 (1992), 179–202.
22. Drazin, R.; Glynn, M.A.; and Kazanjian, R. Multilevel theorizing about creativity in orga-
nizations: A sensemaking perspective. Academy of Management Review, 24, 2 (1999), 286–307.
23. Elam, J., and Mead, M. Can software influence creativity? Information Systems Re-
search, 1, 1 (1990), 1–22.
24. Faraj, S., and Sproull, L. Coordinating expertise in software development teams. Man-
agement Science, 46, 12 (2000), 1554–1568.
25. Fornell, C., and Larcker, D. Structural equation models with unobservable variables and
measurement errors. Journal of Marketing Research, 18, 2 (1981), 39–50.
26. Geisser, S. The predictive sample reuse method with applications. Journal of American
Statistical Association, 70, 2 (1975), 320–328.
27. Grant, R. Prospering in dynamically-competitive environments: Organizational capabil-
ity as knowledge integration. Organization Science, 7, 4 (1996), 375–387.
28. Guinan, P.J.; Cooprider, J.G.; and Faraj, S. Enabling software development team perfor-
mance during requirements definition: A behavioral versus technical approach. Information
Systems Research, 9, 2 (1998), 101–125.
29. Hansen, H. Knowledge networks: Explaining effective knowledge sharing in multiunit
companies. Organization Science, 13, 3 (2002), 232–248.
01 tiwana.pmd 6/15/2005, 11:21 PM39
40 AMRIT TIWANA AND EPHRAIM R. MCLEAN
30. Hargadon, A., and Sutton, R. Technology brokering and innovation in a product devel-
opment firm. Administrative Science Quarterly, 42, 4 (1997), 716–749.
31. Henderson, J.C., and Lee, S. Managing I/S design teams: A control theories perspective.
Management Science, 38, 6 (1992), 757–777.
32. Hickey, A., and Davis, A. A unified model of requirements elicitation. Journal of Man-
agement Information Systems, 20, 4 (Spring 2004), 65–84.
33. Hulland, J. Use of partial least squares in strategic management research: A review of
four recent studies. Strategic Management Journal, 20, 2 (1999), 195–204.
34. Iivari, J.; Hirschheim, R.; and Klein, H. A dynamic framework for classifying informa-
tion systems development methodologies and approaches. Journal of Management Informa-
tion Systems, 17, 3 (Winter 2000–2001), 179–218.
35. Kale, P.; Singh, H.; and Perlmutter, H. Learning and protection of proprietary assets in
strategic alliances: Building relational capital. Strategic Management Journal, 21, 3 (2000),
217–237.
36. Klein, K., and Kozlowski, S. From micro to meso: Critical steps in conceptualizing and
conducting multilevel research. Organizational Research Methods, 3, 3 (2000), 211–236.
37. Klein, K.; Tosi, H.; and Nannella, A. Multilevel theory building: Benefits, barriers, and
new developments. Academy of Management Review, 24, 2 (1999), 243–248.
38. Kraut, R., and Streeter, L. Coordination in software development. Communications of
the ACM, 38, 3 (1995), 69–81.
39. Lederer, A.L., and Mendelow, A.L. Coordination of information systems plans with
business plans. Journal of Management Information Systems, 6, 2 (Fall 1986), 5–19.
40. Lovelace, K.; Shapiro, D.; and Weingart, L. Maximizing cross-functional new product
teams’ innovativeness and constraint adherence: A conflict communications perspective. Acad-
emy of Management Journal, 44, 4 (2001), 779–793.
41. MacCrimmon, K., and Wagner, C. Stimulating ideas through software. Management
Science, 40, 11 (1994), 1514–1532.
42. Marakas, G., and Elam, J. Creativity enhancement in problem solving: Through soft-
ware or process? Management Science, 43, 8 (1997), 1136–1146.
43. Massetti, B. An empirical examination of the value of creativity support systems on idea
generation. MIS Quarterly, 20, 1 (March 1996), 83–97.
44. Mohammed, S. Team mental models in a team knowledge framework: Expanding theory
and measurement across disciplinary boundaries. Journal of Organizational Behavior, 22, 2
(2001), 89–106.
45. Molm, L.; Peterson, G.; and Takashaki, N. Power in negotiated and reciprocal exchange.
American Sociological Review, 64, 6 (1999), 876–890.
46. Mumford, M., and Gustafson, S. Creativity syndrome: Integration, application, and in-
novation. Psychological Bulletin, 103, 1 (1988), 27–43.
47. Nidumolu, S., and Subramani, M. The matrix of control: Combining process and struc-
ture approaches to managing software development. Journal of Management Information Sys-
tems, 20, 3 (Winter 2003–4), 159–196.
48. Nunally, J. Psychometric Theory. New York: McGraw-Hill, 1978.
49. Nunamaker, J.; Applegate, L.; and Konsynski, B. Facilitating group creativity: Experi-
ence with a group decision support system. Journal of Management Information Systems, 3, 4
(Spring 1987), 6–19.
50. Nunamaker, J.; Chen, M.; and Purdin, T. Systems development in information systems
research. Journal of Management Information Systems, 7, 3 (Winter 1990–91), 89–106.
51. Ocker, R.; Hiltz, S.R.; Turoff, M.; and Fjermestad, J. The effects of distributed groups
support and process restructuring on software requirements development teams: Results on
creativity and quality. Journal of Management Information Systems, 12, 3 (Winter 1995–96),
127–153.
52. Okhuysen, G., and Eisenhardt, K. Integrating knowledge in groups: How formal inter-
ventions enable flexibility. Organization Science, 13, 4 (2002), 370–386.
53. Perry-Smith, J., and Shalley, C. The social side of creativity: A static and dynamic social
network perspective. Academy of Management Review, 28, 1 (2003), 89–106.
54. Poppo, L., and Zenger, T. Testing alternative theories of the firm: Transaction cost,
knowledge-based, and measurement explanations for make-or-buy decisions in information
services. Strategic Management Journal, 19, 9 (1998), 853–877.
01 tiwana.pmd 6/15/2005, 11:21 PM40
EXPERTISE INTEGRATION AND CREATIVITY IN INFORMATION SYSTEMS DEVELOPMENT 41
55. Robillard, P. The role of knowledge in software development. Communications of the
ACM, 42, 1 (1999), 87–92.
56. Rulke, D., and Galaskiewicz, J. Distribution of knowledge, group network structure, and
group performance. Management Science, 46, 5 (2000), 612–625.
57. Rus, I., and Lindvall, M. Knowledge management in software engineering. IEEE Soft-
ware, 19, 3 (2002), 26–38.
58. Satzinger, J.W.; Garfield, M.J.; and Nagasundaram, M. The creative process: The effects
of group memory on individual idea generation. Journal of Management Information Systems,
15, 4 (Spring 1999), 143–160.
59. Schniderman, B. Creativity support tools. Communications of the ACM, 45, 10 (2002),
116–120.
60. Sethi, R.; Smith, D.; and Park, W. Cross-functional product development teams, creativ-
ity, and the innovativeness of new consumer products. Journal of Marketing Research, 38, 1
(2001), 73–85.
61. Stein, E., and Vandenbosch, B. Organizational learning during advanced stages of sys-
tem development: Opportunities and obstacles. Journal of Management Information Systems,
13, 2 (Fall 1996), 115–136.
62. Subramanian, G., and Zarnich, G. An examination of some software development effort
and productivity determinants in ICASE tool projects. Journal of Management Information
Systems, 12, 4 (Spring 1996), 143–160.
63. Sutton, R., and Hargadon, A. Brainstorming groups in context: Effectiveness in a prod-
uct design firm. Administrative Science Quarterly, 41, 4 (1996), 685–718.
64. Szulanski, G. Exploring internal stickiness: Impediments to the transfer of best practice
within the firm. Strategic Management Journal, 17, 2 (1996), 27–43.
65. Tuttle, B. A study of staff turnover, acquisition, and assimilation and their impact on
software development cost and schedule. Journal of Management Information Systems, 6, 1
(Summer 1989), 21–40.
66. Unsworth, K. Unpacking creativity. Academy of Management Review, 26, 2 (2001),
289–297.
67. Van den Bosch, F.; Volberda, H.; and Boer, M. Coevolution of firm absorptive capacity
and knowledge environment: Organizational forms and combinative capabilities. Organiza-
tion Science, 10, 5 (1999), 551–568.
68. Van Der Zee, J., and De Jong, B. Alignment alone is not enough: Integrating business
and information technology management with the balanced business scorecard. Journal of
Management Information Systems, 16, 2 (Fall 1999), 137–156.
69. Vessey, I., and Conger, S. Learning to specify information requirements: The relation-
ship between application and methodology. Journal of Management Information Systems, 10,
2 (Fall 1993), 177–201.
70. Walz, D.; Elam, J.; and Curtis, B. Inside a software design team: Knowledge, sharing,
and integration. Communications of the ACM, 36, 10 (1993), 63–77.
71. Watson, R.; Kelly, G.; Galliers, R.; and Brancheau, J. Key issues in information systems
management: An international perspective. Journal of Management Information Systems, 13,
4 (Spring 1997), 91–115.
72. Weick, K. The Social Psychology of Organizing. Reading, MA: McGraw-Hill, 1979.
73. Wierenga, B., and van Bruggen, G.H. The dependent variable in research into the effects
of creativity support systems: Quality and quantity of ideas. MIS Quarterly, 22, 1 (1998),
81–87.
74. Woodman, R.; Sawyer, J.; and Griffin, R. Toward a theory of organizational creativity.
Academy of Management Review, 18, 2 (1993), 293–321.
75. Zahra, S., and George, G. Absorptive capacity: A review, reconceptualization and exten-
sion. Academy of Management Review, 27, 2 (2002), 185–203.
01 tiwana.pmd 6/15/2005, 11:21 PM41
42 AMRIT TIWANA AND EPHRAIM R. MCLEAN
Appendix. Scale Items and Results of Factor Analysis
Factorb
Item Scale itemaCARHKS
Creativity
C1Our team frequently experiments with alternative ways to carry out our work. 0.757 0.155 0.328 –0.005 0.153 –0.001
C2Our team is highly imaginative in thinking about new or better ways to perform
our tasks. 0.623 0.207 0.226 –0.003 0.009 –0.001
C3When a nonroutine matter comes up in our work, we often invent new ways to
handle the situation. 0.519 0.392 0.468 –0.102 0.274 –0.002
Absorptive capacity
A1Overall, members of this team can interrelate to each other’s unique skills
and abilities. 0.284 0.731 0.461 –0.008 0.189 0.000
A2Overall, members of this team can interrelate to each other’s unique expertise. 0.225 0.837 0.477 –0.007 0.122 –0.001
A3Members of this team recognize the potential value of their peers’ expertise. 0.321 0.533 0.453 –0.009 0.235 0.000
Relational capital
R1There is close, personal interaction among team members at multiple levels. 0.276 0.174 0.689 0.226 0.003 0.001
R2At multiple levels, this project team is characterized by mutual respect among
members. 0.114 0.270 0.839 0.136 0.153 0.152
R3At multiple levels, this project team is characterized by mutual trust among
members. 0.119 0.243 0.841 0.151 0.153 0.121
R4At multiple levels, this project team is characterized by personal friendship
between members. 0.197 0.120 0.565 0.150 0.104 –0.001
R5At multiple levels, this project team is characterized by high reciprocity among
members. 0.198 0.180 0.754 0.195 0.154 0.005
01 tiwana.pmd 6/15/2005, 11:21 PM42
EXPERTISE INTEGRATION AND CREATIVITY IN INFORMATION SYSTEMS DEVELOPMENT 43
Expertise heterogeneity
H1Members of this team vary widely in their areas of expertise. 0.003 0.001 0.009 0.892 0.008 0.173
H2Members of this team have a variety of different backgrounds and experiences. –0.005 –0.003 0.194 0.918 0.130 0.004
H3Members of this team have skills and abilities that complement each other’s. –0.004 –0.005 0.218 0.917 0.116 0.004
Expertise integration
K1Members of this team synthesize and integrate their individual expertise at
the project level. 0.329 0.299 0.419 0.191 0.454 0.002
K2Members of this team span several areas of exper tise to develop shared
project concepts. 0.328 0.381 0.414 0.269 0.457 0.122
K3Members of this team can clearly see how different pieces of this project
fit together. 0.277 0.255 0.414 0.294 0.693 0.005
K4Members of this team competently blend new project-related knowledge with
what they already know. 0.269 0.301 0.486 0.273 0.567 0.102
Project success (manager’s evaluation)
S1In light of marketplace-mandated changes and new business requirements
that arose
during
project execution, at the present time, this project delivers
all
desirable features and functionality. –0.001 –0.003 0.008 0.004 0.000 0.884
S2In light of marketplace-mandated changes and new business requirements
that arose
during
project execution, at the present time, this project meets
key project objectives and business needs. 0.000 0.003 0.004 0.170 0.001 0.716
S3In light of marketplace-mandated changes and new business requirements
that arose
during
project execution, at the present time, this project overall
is very successful. 0.000 0.002 0.002 0.133 0.006 0.874
Eigenvalue 6.11 3.32 2.85 2.34 1.10 0.94
Percent variance explained 26.6 14.4 12.4 10.2 4.7 4.1
Notes: a Scale range: 1 = Strongly disagree; 5 = Strongly agree; b Exploratory factor analysis used the maximum likelihood procedure. Loadings in boldface indicate
factor structures.
01 tiwana.pmd 6/15/2005, 11:21 PM43
