Access to this full-text is provided by American Psychological Association.
Content available from Journal of Applied Psychology
This content is subject to copyright. Terms and conditions apply.
Navigating Inter-Team Competition: How Information Broker Teams
Achieve Team Innovation
Thomas Taiyi Yan
1
, Vijaya Venkataramani
2
, Chaoying Tang
3
, and Giles Hirst
4
1
Department of Organisation and Innovation, School of Management, University College London
2
Department of Management and Organization, Robert H. Smith School of Business, University of Maryland, College Park
3
Department of Business Management, School of Economics and Management, University of Chinese Academy of Sciences
4
Research School of Management, ANU College of Business and Economics, Australian National University
Organizations are increasingly using teams to stimulate innovation. Often, these teams share knowledge and
information with each other to help achieve their goals, while also competing for resources and striving to
outperform each other. Importantly, based on their industry, the nature of work, or prior history, some teams
may face more competition from peer teams than others. Our research examines how teams’competitive
relations with other teams in the organization operate in tandem with their collaborative inter-team
information exchange relations in impacting their innovation. Using two studies—afield study of 73
knowledge-intensive teams in high-tech engineering firms and a team-based network experimental study of
162 teams—we find that a high degree of overall competition with many peer teams reduces a focal team’s
ability to acquire and utilize diverse knowledge from these teams (i.e., inter-team knowledge integration),
thereby hindering team innovation. However, applying insights from network structural hole theory, we find
that when a focal team occupies a brokerage position in the inter-team information exchange network, this
can help buffer the effects of competition in getting access to knowledge resources from other teams, thus
enabling their innovation. Additionally, we find that focal broker teams’dealmaking and network
obstruction behaviors explain these effects.
Keywords: team innovation, inter-team competition, brokerage, social networks
As innovation is crucial for organizational survival and
competitiveness, organizations are increasingly using dedicated
innovation teams to leverage the diversity of employees’skills and
expertise (van Knippenberg, 2017;Wuchty et al., 2007). Given the
interconnected nature of organizational work (Hackman & Katz,
2010), these teams often benefit a great deal from interacting with
each other and informally exchanging new knowledge and ideas,
which provide critical stimuli for innovation (Reagans & McEvily,
2003). At the same time, the practitioner literature also provides
ample evidence of intense competition between innovation teams
in organizations such as Tencent, Netflix, and Amazon where
teams actively compete in outperforming one another (Rathi, 2014).
Thus, organizational teams are often embedded in an ecosystem
of collaborative and competitive between-team interactions (Tsai,
2002). This raises interesting questions: Can competitive relations
among teams, while motivating them to innovate, also undermine
the informal inter-team interactions that are crucial to exchange
knowledge and innovate (C. Chen & Deng, 2018)? How then do
teams navigate this tension to learn and benefit from new knowledge
developed by peer teams?
In addressing these questions, we borrow from group social
capital theory (Oh et al., 2006), which suggests that teams in
organizations are embedded in networks of various types of
interactions with other peer teams and that the nature and structure of
This article was published Online First August 29, 2024.
Bryan D. Edwards served as action editor.
Thomas Taiyi Yan https://orcid.org/0000-0002-1327-8664
This research was in part supported by the National Natural Science
Foundation of China (Grants 71932009, 71974178) awarded to Chaoying
Tang. Earlier versions have been presented at Academy of Management
Annual Meeting.
The authors thank Inga Hoever for sharing experimental materials; Joshua
Becker and Chris To for help with Empirica; Sruthi Thatchenkery, Gavin
Kilduff, and Joe Labianca for helpful advice; Sam Dupret, Seth Fletcher,
Andrei Viziteu, Jo Burr, and Zillur Rahman for research assistance; and the
Maryland Behavioral Lab and University College London Reading Group
for their valuable input. Thomas Taiyi Yan thanks Nia Tang, Nancy Zhang,
and Greg Yan for their indispensable support.
Open Access funding provided by University of Chinese Academy of Science:
This work is licensed under a Creative Commons Attribution 4.0 International
License (CC BY 4.0; https://creativecommons.org/licenses/by/4.0). This license
permits copying and redistributing the work in any medium or format, as well as
adapting the material for any purpose, even commercially.
Thomas Taiyi Yan played a lead role in conceptualization, data curation,
formal analysis, investigation, methodology, project administration, software,
supervision, validation, visualization, writing–original draft, and writing–
review and editing and an equal role in funding acquisition. Vijaya
Venkataramani played a supporting role in conceptualization, methodology,
supervision, writing–original draft,and writing–review and editing. Chaoying
Tang played a supporting role in data curation and funding acquisition. Giles
Hirst played a supporting role in writing–review and editing.
Correspondence concerning this article should be addressed to
Chaoying Tang, Department of Business Management, School of
Economics and Management, University of Chinese Academy of
Sciences, No. 80, Zhongguancun East Road, Haidian District, Beijing
100049, China. Email: tcy@ucas.ac.cn
Journal of Applied Psychology
© 2024 The Author(s) 2025, Vol. 110, No. 1, 27–48
ISSN: 0021-9010 https://doi.org/10.1037/apl0001216
27
these between-team dyadic interactions provide opportunities as
well as exert constraints in influencing a focal team’s outcomes
(Borgatti et al., 2009;Oh et al., 2006). Guided by this perspective,
we first examine the network of inter-team competition ties in
organizations (e.g., To et al., 2020;Tsai, 2002) and how it affects
teams’ability to innovate. Whether due to their prior history or
nature of work, teams in organizations often develop unique
competitive relations with some target teams more than others and
specifically strive to outperform them (Garcia et al., 2013;To et al.,
2020). As a result, even within the same organizational setting,
teams can face varying degrees of overall inter-team competition,
depending on how many other teams view a focal team as their
competitor. Building on this premise, we demonstrate how teams
facing a high degree of between-team dyadic competition are
constrained in their innovation efforts because this competition
undermines their ability to acquire and utilize diverse knowledge
resources from peer teams (i.e., inter-team knowledge integration)
that are critical for innovation.
That said, there is also significant “variation in competitive success”
(Burt, 1992, p. 5), where some teams, despite the competition they
face, manage to innovate successfully. Group social capital theory (Oh
et al., 2006) offers a possible explanation and suggests that constraints
imposed by one type of network can potentially be overcome by a
team’s advantage in another type of network. For example, in addition
to having competitive ties, teams also have other types of ties with
each other such as informal information exchange. Borrowing from
structural hole theory, we argue that while competitive ties with
specific teams might restrict the amount of knowledge a focal team
receives from these teams directly, teams that occupy a strategic
brokerage position in the overall network of information exchange
(i.e., act as “bridges”between unconnected teams; Burt, 1992) can
better overcome such competition to innovate. Bridging structural
holes between teams in the critical flow of information allows a focal
team to not only gain alterative means to access knowledge that
might be cut off by competitors but also allows it to exert control
over the flow of information to competitors (Burt, 1992;Soda
et al., 2018).
In examining these relationships, we make several important
contributions. At a broad level, we use a social network perspective to
highlight that team innovation emerges within a broader ecosystem of
informal competitive and collaborative network relations with other
teams. Although prior work has examined factors such as team
composition (Baer et al., 2014), team expertise diversity (van
Knippenberg & Mell, 2016), psychological safety and authentic
emotional climates (Edmondson, 1999;Parke et al., 2022), and internal
team networks (e.g., Venkataramani & Tang, 2024), there is scant
research on how different types of inter-team relations in organizations
can impact team innovation. This is an important oversight because
between-team ties serve as a crucial source for importing relevant ideas
or information into the team domain (e.g., Hansen, 1999;Tsai, 2001)
and facilitate innovation. At the same time, organizational teams also
have competitive relations with each other (To et al., 2020) and are
therefore motivated to thwart each other’s progress in achieving goals
as well. Without studying this ecosystem of between-team ties and how
different types of inter-team interactions may jointly affect their
outcomes, our understanding of team innovation would be incomplete
(Labianca & Brass, 2006;Shipilov et al., 2014).
Second, and related to the above, the current article contributes to
an important conversation in the inter-team competition literature.
While previous work has provided valuable insights by identifying
different types of structural interdependence (e.g., incentives such as
zero-sum contests where the winner takes all or means-based
interdependence; Stanne et al., 1999) as drivers of competition, this
approach fails to recognize that competition is not just a response to
common structural incentives but is also influenced by dynamics
that are unique to specific pairs of teams (see To et al., 2020, for a
review). For example, even if there is no structural incentive to
compete, two teams may still compete intensely because of dyadic
factors such as their prior history, psychology, or the unique nature
of their work (Garcia et al., 2013;Porac & Thomas, 1990).
Similarly, such dyadic factors can also lead teams to choose not
to compete despite the existence of structural incentives (e.g.,
Jarzabkowski & Bednarek, 2018). Therefore, by examining
competition dyadically between teams in terms of their overall
striving to outperform specific target teams due to various reasons,
rather than more narrowly due to the nature of their structural
outcome or means interdependence (Stanne et al., 1999), the current
research contributes to a more holistic understanding of the intricate
competitive dynamics among organizational teams.
Third, we provide unique insights regarding the boundary
conditions of inter-team competition. In this regard, we highlight
that constraints posed by one type of between-team network
interaction such as competition can be mitigated by advantages
arising from other kinds of network interactions. We show how
teams that occupy information brokerage positions effectively stave
off the knowledge deficit brought on by inter-team competition and
continue to integrate knowledge from peer teams and innovate. On
the other hand, without bridging structural holes in the inter-team
information network, teams that encounter a high degree of
competition suffer a “double whammy”—not only do they already
lack access to useful knowledge resources due to their nonbroker
positions, but whatever knowledge supply is available to them is also
significantly more vulnerable to the adverse effects of competition.
In exploring this, we offer valuable insights into the strategic
choices and behaviors of broker teams when facing competition.
When facing a high degree of competition, a focal team that has
an information brokerage advantage engages more in behaviors
such as dealmaking (creating quid pro quo bargains with alters)
and obstruction (i.e., controlling or preventing information flow
to certain alters in the network). Thus, our research contributes
valuable microbehavioral evidence into how information brokerage
motivates teams to engage in strategic behaviors to combat the
adversity brought on by inter-team competition (Halevy et al., 2020;
Tasselli et al., 2015).
Theoretical Development and Hypotheses
Organizational teams are embedded within a broader social
structure of formal and informal interactions with other teams.
Drawing on a network theoretic tradition of focusing on the
configuration of such interactions, group social capital theory (Oh et
al., 2004,2006) proposes that teams’pattern of interactions with
peer teams and their structural positions in the networks of such
between-team interactions can provide important opportunities for,
as well as impose constraints on, team outcomes (Ancona, 1990;
Tsai, 2001). Guided by group social capital theory, we focus on two
types of interactions that teams have with one another—competitive
ties and collaborative information exchange ties—and examine how
28 YAN, VENKATARAMANI, TANG, AND HIRST
the pattern of these ties and teams’position within these networks
affects their ability to innovate. Specifically, we propose that while
teams’competitive ties may constrain their ability to innovate, their
strategic position in the between-team information sharing network
can buffer such effects.
Inter-Team Competition: Dyadic Competition
Between Teams
Historically, inter-team competition has been studied according to
the lens of social interdependence theory (Deutsch, 1949;D. W.
Johnson & Johnson, 2002). Social interdependence exists when
individuals or teams share common goals and each team’s outcomes
are affected by the actions of other teams (D. Johnson & Johnson,
1989). The basic premise of social interdependence theory is that
the way goals and rewards are structured determines how actors
(i.e., individuals or teams) interact, and in turn, the outcomes of the
situation. Accordingly, D. Johnson and Johnson (1989) identified
types of competition based on outcome versus means interdepen-
dence. Outcome interdependence specifies how the goals and
rewards that actors strive to achieve are related. Means interdepen-
dence specifies the actions required on the part of participants to
achieve their goals, and it exists when a task is structured such that
two or more actors are required to jointly complete it. Whereas some
amount of negative outcome interdependence (e.g., where a goal
such as a reward or promotion can only be achieved by one or few
actors and where one actor’s success reduces the chances of success
of other actors) may exist in competitions, some competitions may
also involve means interdependence (e.g., a chess game; D. W.
Johnson & Johnson, 2002, p. 124). In this regard, Stanne et al.
(1999) found in an individual-level meta-analysis that competition,
operationalized as negative structural outcome interdependence,
was positively related to individual performance on motor skill tasks
only when there was no means interdependence. Sherif’s (1956)
anecdotal research on competition and conflict in boys’summer
camps also suggested similar findings.
1
The structural outcome interdependence perspective has been the
most prevalent framework in existing work on competition (D. W.
Johnson & Johnson, 2002;Kistruck et al., 2016). Therefore, extant
work has predominantly operationalized competition between teams
in terms of a “competitive social situation”in which a zero-sum
outcome (e.g., a prize, bonus) can only be achieved by one or few
teams (e.g., Baer et al., 2010,2014;Boudreau et al., 2011). As a
result, in a given situation (i.e., an organization), social interdepen-
dence theory presumes that all teams under the same structural
competitive incentive compete equally with one another in winning
the prize or reward. Yet, recent research has indicated that this
structural conceptualization of competition may be insufficient to
describe how organizational teams actually experience inter-team
competition (Eisenkraft et al., 2017;To et al., 2020). Although
organizations may have common structural incentives (e.g., annual
performance ranking of teams for bonus allocations) that may give
all teams a baseline motivation to achieve higher performance,
teams often also develop highly differentiated competitive ties
with peer teams wherein they experience greater competition with
(i.e., strive to outperform) specific target teams but not others (Ku
et al., 2005;Porac & Thomas, 1994;Shah, 1998).
Building on this work that highlights how competition can exist
at the dyadic level, To et al. (2020) conceptualized competition as
a focal actor’s“striving to outperform a specific target.”Thus,
competition between two teams exists when “a focal team strives to
outperform a specific target team”(p. 911). Such competition with
specific peer teams could arise due to various factors such as a focal
and target team vying for the same scarce resources (i.e., dyadic
outcome interdependence), similarity in task domains (Tesser &
Smith, 1980), past history and experiences with one another (Ku
et al., 2005;Lount & Phillips, 2007), or a combination of these
factors. Indeed, dyadic competition could well exist even when there
is no structural competition, based on other factors such as shared
history and similarity (Garcia et al., 2013;Porac & Thomas, 1994).
Similarly, despite the existence of structural incentives, teams may
well choose to not compete with each other due to unique dyadic
reasons (Jarzabkowski & Bednarek, 2018).
To illustrate, Kilduff et al. (2010) studied teams within the Pac-10
National Collegiate Athletic Association basketball division and
found that even under the same structural competitive situation,
specific pairs of teams had unique competitive relationships, which
were predicted by their idiosyncratic histories with one another,
while others did not. For example, whereas Oregon State and
University of Oregon had a fierce competition with each other
and were motivated to specifically outperform one another, Oregon
State had comparatively lesser competition with the other eight
teams. In fact, these authors found that 50% of variance in the
competition ratings among these teams were attributed to between-
team dyadic differences. Similarly, Hansen et al. (2005) showed
that research and development teams within the same firm had
competitive ties with some teams but not others.
This is the starting point of our article—the prevalence of
dyadically differentiated competition ties between specific teams in
organizations and, as group social capital theory posits (Oh et al.,
2006), the inter-team competition network comprising of such teams
and the ties (or lack thereof) between them. As a result of the
difference in the number and intensity of dyadic competitive ties that
each team encounters in the network, there is significant variation
in the aggregate, overall degree of inter-team competition
experienced by each team, despite being part of the same context.
For instance, in the Pac-10 example above, University of Arizona
encountered a much higher overall degree of competition from peer
teams (an average rating of 7.47 from the other nine teams on a scale
of 1–10 of competitive intensity) than Oregon State (average rating
of 3.38; Kilduff et al., 2010). This overall degree of competition
experienced by a focal team—defined as the number and intensity of
dyadic competitive ties that a focal team has with other peer teams,
aggregated across these teams—is the focus of our research.
1
It is important to note that in this study, we focus on teams that have well-
defined team membership and strive autonomously toward their own
innovation-focused goals. In other words, while they may often interact
informally to exchange resources and knowledge, there is no formal means
interdependence between these teams where a team’s own work requires
input from another team without which it cannot be completed (Stanne et al.,
1999). Examples of such autonomous teams include software development
teams, creative design teams (e.g. advertising teams), and professional
service teams (e.g. consultant teams; Oh et al., 2006). In this sense, we do not
focus on multiteam systems, such as a military strike unit where a number of
teams are formally designated to coordinate (i.e., have means interdepen-
dence) to achieve a superordinate goal (e.g. Davison et al., 2012). In such
cases, formal means interdependence among teams might play a role in
qualifying our proposed relationships.
INTER-TEAM COMPETITION AND INNOVATION 29
Degree of Overall Inter-Team Competition and
Team Innovation
Experiencing competition can significantly influence focal teams’
psychology and actions (M. Chen et al., 2007;Garcia et al., 2013).
For example, prior research finds that inter-team competition acts as
an external threat and creates a stronger bond among team members
internally (Halevy et al., 2008), enhances members’intrinsic
motivation (Cikara et al., 2011), reduces inefficiency and free-riding
(Erev et al., 1993), and facilitates intrateam collaboration and
coordination (Baer et al., 2010). However, this work, predominantly
conducted in lab settings and focusing mainly on within-team
processes, invariably treats teams as standalone, independent
entities, thus overlooking the crucial between-team interactions
that also occur (see Sherif, 1956;Tsai, 2001, for notable exceptions).
Organizational teams, while operating autonomously toward
their own goals, often interact with one another informally to gather
unique information and resources, as well as learn and benefit from
new ideas developed by each other (Ancona & Cladwell, 1992b;Oh
et al., 2006;Reagans & Zuckerman, 2001). Thus, such interactions
help teams achieve inter-team knowledge integration,defined as a
focal team’s acquisition of knowledge from peer teams in the
organization (Gupta & Govindarajan, 2000), and the utilization of
this knowledge in its own work (Reagans & McEvily, 2003).
Because it often provides nonredundant knowledge and perspec-
tives beyond the internal knowledge base available to the focal
team, inter-team knowledge integration is particularly beneficial
for innovation (Anderson et al., 2014;Phelps et al., 2012;van
Knippenberg, 2017). However, competition between teams is likely
to affect it.
We suggest that when a focal team faces a high degree of inter-
team competition (i.e., has a large number of between-team dyadic
competitive ties), it adversely impacts the focal team’s ability to
acquire and utilize knowledge from them. First, it is likely to acquire
lesser amounts of overall information directly from competing peer
teams due to their reluctance to facilitate the focal team’s progress.
Second, even when competing teams share some knowledge, this
knowledge might be incomplete, inaccurate, or even misleading. As
prior studies (albeit at the individual level of analysis) suggest,
actors are motivated to withhold, conceal, and misrepresent valuable
knowledge from each other if this knowledge can help their
competitors’performance (Connelly et al., 2012;Garcia et al., 2010;
Reh et al., 2018;Steinel & De Dreu, 2004). In fact, such antagonistic
tendencies have been shown to be more pronounced at the team
level as compared to the individual level (Wildschut et al., 2003).
Along these lines, Hansen et al. (2005) found that competition made
focal teams’knowledge search more difficult and time-consuming.
Finally, even if competing peer teams provide some knowledge,
the effective utilization of this knowledge by the focal team is
likely to be compromised. Competition has been shown to lead
to suspicion about the underlying motivation of the knowledge
provider and the potential truthfulness and quality of the knowledge
in question, which in turn diminishes a focal team’s motivation
to utilize this knowledge (Menon et al., 2006). Also, a focal team
might be reluctant to utilize knowledge acquired from a competitor
because doing so might indicate an admission of incompetence
(Gupta & Govindarajan, 2000). Thus, we hypothesize:
Hypothesis 1: The degree of inter-team competition encoun-
tered by a focal team is negatively associated with its inter-team
knowledge integration (i.e., its acquisition and utilization of
knowledge from peer teams in the organization).
Next, we propose that a focal team’s inter-team knowledge
integration is positively associated with team innovation. Team
innovation is defined as the process, outcomes, and products of
attempts to develop and introduce new and improved ways of doing
things within the team (Anderson et al., 2014). Thus, innovation
consists of two aspects, idea generation as well as its implementa-
tion, and a focal team’s ability to acquire and utilize knowledge from
peer teams in the organization is particularly important for both.
First, knowledge acquired from external sources tends to be
different or nonredundant from what the focal team already knows
and thus especially likely to facilitate new perspectives, challenge
the status quo, and increase the likelihood of developing novel
solutions (Phelps et al., 2012;van Knippenberg, 2017). Ancona and
Caldwell (1992a) showed that performance of consulting teams
benefited significantly when they could acquire knowledge from
outside sources. Similarly, other research has found that knowledge
acquired from external sources tends to be diverse and novel
and, when used in the focal team’s work, tends to challenge its
existing cognitive schemas, prompting them to think differently
and stimulating innovation (Hargadon & Sutton, 1997;Reagans &
Zuckerman, 2001).
Second, innovation occurs not only when a focal team integrates
ideas and information from diverse peer teams to create new
products and services but also when it recognizes analogies between
its own situation and those experienced by others and adapts their
solutions to the current situation. Thus, access to information about
potential constraints, problems, and challenges that peer teams
experience is also important for the implementation of these new
ideas. Taking these arguments together, we propose that teams that
are capable of inter-team knowledge integration are more likely to
be more innovative (van Knippenberg, 2017).
Hypothesis 2: A focal team’s inter-team knowledge integration
is positively associated with team innovation.
Hypothesis 3: inter-team knowledge integration mediates the
negative relationship between the degree of inter-team
competition faced by a focal team and team innovation.
Variations in Competitive Success: The Role of Inter-
Team Information Brokerage
Although facing a high degree of inter-team competition can
adversely impact a focal team’s ability to integrate knowledge
and, thus, its innovation, some teams are still able to overcome the
negative effects of competition to innovate successfully. Prior work
on moderators of competition has examined task characteristics such
as means interdependence (albeit at the individual level; Stanne et
al., 1999) and within-team characteristics such as gender composi-
tion (Baer et al., 2014) and team regulatory focus (Beersma et al.,
2013). However, team studies, while focusing on internal team
processes such as coordination and cohesion, are unable to shed
30 YAN, VENKATARAMANI, TANG, AND HIRST
light on the between-team knowledge integration processes affected
by inter-team competition.
Group social capital theory (Oh et al., 2006) suggests that one
explanation may lie in the fact that teams are embedded in different
types of networks and that the structural position that a focal team
occupies in one type of network can provide unique benefits that
might offset the disadvantages in another (e.g., Venkataramani et al.,
2013). For example, beyond competitive ties, teams are also
connected via informal ties of communication and information
exchange with each other. While having a competitive tie with a
target team might suggest a lack of information exchange tie with
that team, this is not necessarily the case (Labianca, 2014). In fact,
competing teams can and sometimes do also have positive
interactions with one another (Tsai, 2001,2002) but may differ
in the strength of these interactions, such as the amount and type of
knowledge they share or the extent to which they collaborate. Thus,
two competing teams could have an information exchange tie, but
only exchange minimal knowledge or low-value information (or
even misleading information; Garcia et al., 2010;Steinel & De Dreu,
2004). Important in this context, however, is that these teams have
ties with other teams in the network as well. We argue that the
broader pattern of information exchange ties that a focal team has
with these other teams has crucial implications for its success in
overcoming competition.
Structural hole theory (Burt, 1992;Kwon et al., 2020) is a natural
fit to shed light on these issues because it was originally formulated
to explain variations in actors’performance success, particularly in
competitive, knowledge-intensive contexts. It argues that actors
that bridge structural holes—that is, act as a broker in connecting
otherwise disconnected nodes in a network (Burt, 1992)—have
unique advantages that can translate to important outcomes such
as performance and innovation (Burt, 2001,2004;Fleming &
Waguespack, 2007). Connecting with different actors that do not
interact with each other provides access to distant social worlds
and, thus, to a wider variety of, as well as less redundant, knowledge.
Moreover, because they are the lone “bridge”between their alters,
brokers have control over when and what knowledge gets transferred
between them, thus providing unique first-mover advantages and
discretionary control. Accordingly, we propose that occupying
brokerage positions in the information exchange network among
peer teams would especially benefit a focal team that is facing a
high degree of competition.
It is important to note that experiencing a high degree of
competition from many other teams and being an information broker
are theoretically and empirically distinct constructs. Given that
competition and information sharing are two different types of ties
that connect teams (e.g., Labianca & Brass, 2006;Venkataramani &
Dalal, 2007), two teams experiencing the same degree of competition
can still have differing patterns of information exchange ties with
other peer teams, thereby occupying information brokerage positions
to varying degrees. Thus, even if a focal team does not receive
needed information from directly competing teams, its brokerage
position can offer strategic advantages. Figure 1 illustrates these
possibilities for a focal Team A.
Figure 1
Illustration of Competition–Information Brokerage Combinations
for Focal Team
A
C
B
D
FG
E
Panel A: High Competition-
High Info Brokerage
A
C
B
D
FG
E
Panel B: High Competition-
Low Info Brokerage
A
C
B
D
FG
E
Panel C: Low Competition-
High Info Brokerage
A
C
B
D
FG
E
Panel D: Low Competition-
Low Info Brokerage
Competition tie Information tie
Note. See the online article for the color version of this figure.
INTER-TEAM COMPETITION AND INNOVATION 31
Panel A illustrates Team A, which encounters a high degree of
competition from multiple teams (Teams B, D, F, G; denoted by
dashed lines), but at the same time, also has information exchange
ties with some of these and other teams (Teams C, E, F, G; denoted
by thickened lines) that are not connected to each other via
information exchange ties. As such, Team A is the only liaison
connecting Team F with Teams G, C, and E; Team G with F, C, and
E; and Team E with F, G, and C, thereby occupying a significant
information brokerage position. In contrast, Panel B illustrates
the case where the focal team (Team A) has the same degree of
competition as in Panel A (i.e., four competitors), but most of its
information exchange partners are also connected with each other,
thereby reducing its brokerage opportunities.
We propose that when encountering a high degree of competition,
being an information broker provides several distinct advantages to
the focal team. First, because they connect otherwise disconnected
teams in the information sharing network, broker teams enjoy
rich information availability (Burt, 1992;Oh et al., 2006). This
advantage is not only manifested in the amount of knowledge that
flows to these teams but also in its novelty and diversity, which
are crucial to innovation (Burt, 2001;van Knippenberg, 2017).
Thus, if a focal team (e.g., Team A in Figure 1, Panel A) is
confronted with competition from other teams, thereby reducing its
direct access to knowledge available from these specific teams,
being a broker in the information network could buffer such effects
by providing alternative sources of knowledge from other alters
(e.g., from Teams E and C as well as being a conduit between them
and others). Similarly, although Team A does not have a direct
information tie with its competitor Team D, it can leverage its
brokerage position to receive such information through Team C,
which has a tie with Team D. On the other hand, when teams do
not possess such positional advantages, inter-team competition is
likely to severely diminish their ability to access knowledge from
other teams.
Second, in addition to information access benefits, brokers also
wield substantial control over knowledge dissemination (i.e., “traffic
controller;”Burt, 1992). Because they serve as the bridge between
disconnected alters, information broker teams can substantially
influence the quantity and quality of information received by
themselves and their alters, which can be a crucial advantage when
facing strong competition. For example, in Panel A, by gaining
access to information that its competitors (say, F and G) may not
have access to, Team A, by virtue of being the broker between
them and other teams such as E, C, or G, may be able to use it as
leverage in trading valued resources from other teams or in getting a
first-user advantage. Furthermore, information broker teams can
exploit the information diversity in their network to make deals
with their information alters (even reluctant ones) to acquire useful
knowledge (e.g., “arbitrage”Soda et al., 2018). For example, Team
A can trade information that it has secured from Team E but only if
Team F provides specific knowledge. Broker teams can also obstruct
knowledge flow to their competitors. Such a subversive form of
network control has received less scholarly attention due to its
subtlety (for exception, see “divisive behavior,”Halevy et al., 2020),
yet it is likely to occur when the focal team encounters intense
competition. For example, Team A (Panel A), which acts as the
bridge between Team E and a competing Team G, can acquire
knowledge from Team E and choose to withhold it from Team G
to deter its progress.
In addition to knowledge acquisition, brokerage also helps teams
to utilize external knowledge more effectively. Brokers, due to their
role as knowledge flow controllers, are responsible for relaying
information between different domains and, therefore, are likely
more experienced and skilled in framing and communicating new
knowledge (Burt, 2004). As such, occupying a brokerage position
equips teams with a “vision advantage”to appraise and translate
knowledge into innovation outcomes (Lingo & O’Mahony, 2010;
Reagans & McEvily, 2003). This ability is likely to be particularly
advantageous when faced with competition wherein the motivation
of certain knowledge providers may not be apparent (Garcia et al.,
2010;Menon et al., 2006). Without the ability to contextualize or
evaluate the heterogenous knowledge they may have access to,
nonbroker teams facing intense competition are especially likely to
be less effective in utilizing their resources in being innovative.
Importantly, broker teams can not only identify their value but can
also build shared vision and consensus around them to translate
them into tangible outcomes (Hülsheger et al., 2009). In summary,
we suggest that while facing a higher degree of inter-team
competition could hinder a focal team’s ability to effectively
integrate knowledge from peer teams, the focal team’s information
brokerage position may mitigate these negative consequences.
Accordingly, we propose,
Hypothesis 4: A focal team’s brokerage in the inter-team
knowledge sharing network moderates the relationship between
inter-team competition and inter-team knowledge integration
such that when brokerage is high, the relationship is less
negative.
Hypothesis 5: The indirect effect (mediation) of inter-team
competition on team innovation via inter-team knowledge
integration (Hypothesis 3) is moderated by the focal team’s
brokerage such that when brokerage is higher, the indirect effect
is less negative.
Overview of Studies
We tested our full theoretical model in two studies. Study 1 was
a network field study sampling 73 engineering teams (employee
N=689) working in technology-related industries in China. Study 2
was a team-based interactive social network experiment (focal team
N=162, total individual N=972) where participants engaged in
a team innovation task. By independently manipulating between-
team dyadic competition and the brokerage position of specific
teams in the inter-team information exchange network, this study
provides causal evidence for the relationships proposed, as well
as rich insights into the focal team’s behavioral strategies.
Transparency and Openness
We describe our sampling plan, all data exclusions (if any),
manipulations, and measures in adherence to the Journal of Applied
Psychology methodological checklist. All analysis codes and
materials are available upon request. Data were analyzed using
UCINET software Version 6.662 (Borgatti et al., 2002), SPSS
Version 27, and Mplus Version 8.4 (Muthén & Muthén, 1998/
2017). We did not preregister these studies’design and analyses.
All studies were approved by the institutional review board at
University of Maryland, College Park (Protocol numbers #1666940;
32 YAN, VENKATARAMANI, TANG, AND HIRST
#1666940-2). Study 2 data and an additional online material and the
appendix are hosted at the Open Science Framework available at
https://tinyurl.com/u7nn9bc4.
Study 1: Method
Sample and Procedures
Guided by past teams research (e.g., Bunderson et al., 2016), we
approached organizations with a set of team research criteria and
received commitments from four engineering firms with a total of
103 project teams.
2
However, due to unexpected work demands, one
of the firms with 26 teams was unable to continue participating
midway. As a result, we could not collect data on our team-level
dependent variable (DV) and thus removed this firm from our
analyses. Among the remaining 77 teams from three organizations,
four teams had missing responses on our dependent variable. Our
final sample thus consisted of 689 employees from 73 engineering
teams nested in three firms. Details such as ownership structure and
technological focus are provided in the additional online material and
the appendix. Given that engineering teams can have multiple
performance goals (e.g., maintenance, troubleshooting), we con-
firmed with firm management that innovation was a salient outcome
for these teams. For instance, the official mission statements of all
these firms prominently featured language related to “innovation.”
In our interactions with firm management, we discovered some
“natural boundaries”(Borgatti et al., 2018) within the second firm
(pseudonymized “Umbrella Tech,”see additional online material,
Table A1). Specifically, the 37 Umbrella Tech teams were organized
in five different product divisions that worked at different locations
and in unrelated industries (e.g., enterprise cloud storage vs.
employee relationship software development). Because of these
natural boundaries, these teams very rarely interacted or competed
with others from another division and firm management advised us
to treat each division as a smaller yet distinct organization and
conduct our survey rollout separately. Therefore, we labeled each
of the five divisions as a “data site”and viewed them as separate
“networks of teams.”Overall, along with the other two firms, there
are in total seven data sites or seven “networks of teams”in our
sample.
All surveys were administered in Mandarin, after following
standard translation and back translation procedures (Brislin, 1976).
Translation from English to Mandarin was undertaken by the first
author, and the back translation was undertaken by the third author.
Discrepancies were resolved by the two authors with input from our
HR liaisons to ensure contextual fit. We then conducted a survey
pretest (Schaffer & Riordan, 2003) with project managers (PMs) and
employees to incorporate any feedback; we administered separate
surveys to team members and project managers at two time points.
At Time 1, employees and project managers provided responses on
our explanatory and control variables. Four weeks later, at Time 2,
project managers rated teams’innovation. To ensure confidentiality,
surveys were distributed to respondents via sealed envelopes and
collected back by the researchers. The average number of teams per
data site was 10.43 (SD =7.7), and the average size of teams was 9.4
members (SD =2.2). Among employees, the average age was 31.3
years (SD =5.4), and 70.0% were male; among project managers,
the average age was 34.1 years (SD =5.1), and 81.7% were male.
All participants had a bachelor’s degree or higher.
Preliminary Interviews
We conducted several interviews with our organizational liaisons
to gather facts about teams’internal operations and between-team
interactions, as well as solicit feedback about our survey design. Our
liaisons include several frontline employees, project managers, as
well as the CEO and head of HR of each company. First, we wanted
to ascertain that organizational teams did in fact experience
between-team dyadic competition with peer teams in a differentiated
manner. To this end, interviews with project managers and the
head of HR first indicated that all teams experienced common
structural incentives such as an end-of-year review and ranking all
teams for allotment of performance bonuses. Thus, this baseline
structural incentive was commonly experienced by all teams. More
importantly, interviews indicated that teams indeed experienced
more competition with specific peer teams than others. For instance,
some managers spoke about their team being competitive against
a specific target team because they perceived the two teams to be
fighting for a funding opportunity or because they perceived each
other operating in similar markets. Other managers spoke of their
team competing against another team because many team members
of these two teams graduated from rival universities or because
of prior interaction history where the two teams had competed for
the same new hires. Oftentimes, respondents also mentioned how
a combination of such factors were at play. These findings affirmed
that teams indeed experienced competition at the dyadic level, and
as a result, some teams encountered a higher overall degree of inter-
team competition than others.
Additionally, we confirmed from our interviewees that teams
were designed to work autonomously to meet respective client
demands. In other words, there was no formal, organizationally
designed means interdependence between teams and they did not
need to coordinate with each other in completing their goals.
However, these teams did engage in informal interactions with one
another (albeit to varying degrees) in sharing information and other
resources. Thus, these interviews ascertained the relevance of this
context for our study.
We also used these interviews to solicit suggestions about which
sources were best suited to provide responses on specific variables.
Given our interest in studying inter-team interactions, both
managers and employees in these interviews indicated that project
managers would be best suited to report their team’s competitive ties
and information exchange ties with peer teams as they were most “in
the thick of things”and were often the ambassador interfacing with
peer teams (Ancona & Caldwell, 1992a). Thus, and also in line with
prior research (Gladstein, 1984;Hansen, 1999;Porac & Thomas,
1994), we used managers as a key informant to measure a focal
team’s competitive ties and information exchange ties. In contrast,
in capturing our mediator (i.e., the extent to which a focal team
acquires and utilizes knowledge from other teams), both suggested
that it was team members who communicated with and acquired
knowledge from members in other teams and were the “end users”
that converted acquired knowledge to tangible innovation outputs.
Hence, we asked all team members to report the team’s inter-team
knowledge integration and aggregated them to the team level.
Finally, our interviews suggested that team managers typically
2
The present research is the first publication from a large data set involved
in a broader research project.
INTER-TEAM COMPETITION AND INNOVATION 33
oversee a number of projects and thus have the ability to discern
different levels of innovativeness. Thus, we used project managers
to measure team innovation, which is also supported by prior work
(e.g., Carnabuci & Dioszegi, 2015).
Measures
We used a network-based approach to measure the extent of
between-team dyadic competition and the extent of information
sharing among teams (e.g., Hansen et al., 2005;Kilduff et al., 2010;
Oh et al., 2004). Consistent with a whole network measurement
approach, project managers were provided with a list of names of all
project teams and their managers in the data site and were asked to
respond to specific questions about them (e.g., Marsden, 1990).
The Degree of Inter-Team Competition Encountered
by the Focal Team (Time 1)
Guided by prior work (e.g., Eisenkraft et al., 2017;Hansen
et al., 2005;To et al., 2020), we measured between-team dyadic
competition ties by asking project managers to respond to the question,
To what extent is your team in competition with this team? Competition
could include situations where you and this team compete for the same
resources, outcomes and support, and where you strive to do better than
this team in terms of your performance and assignments.
Respondents used a 5-point scale (0 =no competition at all to 4 =
very intense competition) to answer the question about every other
team in their data site.
Thus, seven inter-team competition network matrices were
constructed, one for each data site. Using the UCINET software
(Version 6.662; Borgatti et al., 2002), we calculated inter-team
competition in-degree centrality. Mathematically, the in-degree
centrality is calculated by sums of the weights of incoming (i.e.,
rated by peer teams) competition ties to the focal team (Freeman,
1978). Thus, high scores on this measure indicate that the focal team
encounters a high degree of competition from peer teams in each
data site. Scores ranged from 0 (i.e., no team competes against
the focal team) to larger positive values as more teams nominated
the focal team as their competitor.
Focal Team’s Inter-Team Information Brokerage
(Time 1)
For each data site, an inter-team information exchange network
was constructed by asking project managers to respond to the
question (adapted from Reinholt et al., 2011), “How frequently does
your project team receive technical knowledge or project-related
information from this team?”using a 5-point Likert scale from
0(not at all)to4(always). Because the calculation of brokerage
requires binary data input, we dichotomized our information
network based on recommendations by Borgatti and Quintane
(2018) such that it retained the most amount of information from
the original weighted network. Compared to alternative options, a
cutoff value of 2 (occasionally) consistently retained the most
amount of variance between the weighted network and the
dichotomized network (average r=.85) across the seven data
sites. Therefore, values greater than or equal to 2 were recoded as 1
and others were coded as 0. We then used these dichotomized ties to
construct seven inter-team information network matrices, one for
each site.
We measured team brokerage using the constraint measure (Burt,
1992). Constraint scores typically range between 0 and 1, though it
can slightly exceed 1 for small networks (Everett & Borgatti, 2020).
In our data, the constraint scores ranged from 0 to 1.125. High constraint
scores mean ego’s access to few structural holes, therefore lower
brokerage. Thus, we report the negative of constraint as brokerage.
Focal Team’s Inter-Team Knowledge Integration (Time 1)
We measured inter-team knowledge integration by asking all
team members to respond to the following two questions about
their team’s interaction with other teams in general (adapted from
Reinholt et al., 2011):
Please indicate the extent to which your team (a) receives knowledge or
information from team members working in other project teams in your
organization, and (b) uses knowledge or information acquired from
team members working in other project teams in your organization.
Respondents used a 5-point scale ranging from 1 (no or very little
extent)to5(a very large extent). This measure had adequate
reliability (α=.81) and aggregation statistics, median R
wg
=.77;
intraclass correlation coefficient (ICC, 1) =.13, ICC(2) =.57, F=
2.32, p<.01. Although a higher ICC(2) would be ideal, a lower
value does not prevent aggregation if R
wg
is high and group variance
is significant (LeBreton & Senter, 2008). Thus, we aggregated
member ratings in each team to operationalize this construct.
Team Innovation (Time 2)
At Time 2, project managers evaluated team innovation by
responding to a four-item measure from De Dreu and West (2001),
using a 5-point Likert scale, ranging from 1 (strongly disagree)to5
(strongly agree). A sample item is, “team members often implement
new ideas to improve the quality of our products and services”(α=.90).
Control Variables
First, we controlled
3
for the size of each inter-team network
because network size covaries with both the centrality-based
competition measure and structural holes in the network (Burt,
1992;Freeman, 1978) and can affect our results. Second, we
controlled for team size as larger teams tend to have more diverse
expertise among its members and thus are likely to be more
innovative (Stewart, 2006). Third, because teams’knowledge
integration and innovative behavior could change with stage of the
project (Gersick, 1988), we controlled for teams’current state of
project completion (reported by project managers, 1 =25% or
below, 4=75% and above). Next, given consistent evidence
supporting a gender difference in competitiveness (Niederle &
Vesterlund, 2011), we controlled for project managers’gender as
this could potentially impact their reported competition ties.
Similarly, because more experienced managers might face more
competition (Gerber et al., 2017) and might have more expertise in
guiding the team’s innovative direction (Wu et al., 2005), we also
3
Our results are robust without control variables, except for network size
because it covaries with centrality-based and structural hole based measures
(e.g. Burt, 2004).
34 YAN, VENKATARAMANI, TANG, AND HIRST
controlled for project manager’s tenure. As prior work has shown
that competition can positively affect within-team cohesion in
facilitating innovation (Baer et al., 2010), we controlled for
members’team identification using a five-item scale by Mael and
Ashforth (1992). Finally, because team task characteristics might
impact teams’reliance on external knowledge (van Knippenberg,
2017), we controlled for task complexity (three-item scale from
Campion et al., 1993) and team task interdependence (three-item
scale from Dean & Snell, 1991), both reported by project managers.
Analytical Strategy
While our data is characterized by a four-level structure (i.e.,
individual, team, data site, organizations), our hypotheses are focused
on between-team differences. Given that the number of data sites (i.e.,
clusters in the data) was quite low (i.e., seven), in order to still account
for nonindependence between teams in a given data site, we
conducted our analysis with cluster-robust standard errors (McNeish
& Stapleton, 2016) in Mplus Version 7.4 (Muthén & Muthén, 1998/
2017). To account for potential differences due to organizations and
data sites, we also created two firm dummy variables (for three firms)
and six data site dummy variables (for seven data sites) and entered
them as fixed effects. However, this did not affect our results in any
way. Following convention, all explanatory variables were grand-
mean centered (Enders & Tofighi, 2007).
Study 1: Results
Table 1 reports descriptive statistics and correlations among all
study variables. When examining responses of slight competition
and above, the average density of inter-team competition networks
across the seven data sites was .34, indicating that inter-team
competition was a tangible phenomenon. Similarly, the average
density of the inter-team information networks was .46, indicating
that inter-team information exchange was frequent.
4
The between-
team dyadic correlation between competitive ties and information
sharing ties was modest (r=.17), suggesting that competing teams
do still share some information with one another (Table 2).
Hypothesis 1, which proposed that the extent of inter-team
competition experienced by a focal team would be negatively related
to its inter-team knowledge integration, was supported (b=−.03,
β=−.43, p<.01). As predicted by Hypothesis 2, inter-team
knowledge integration was positively associated with team
innovation (b=.37, β=.18, p<.01). Supporting Hypothesis 3,
our mediation analysis using a Monte Carlo method with 20,000
iterations (Slig & Preacher, 2008) showed a significant uncondi-
tional indirect effect of inter-team competition on team innovation
via inter-team knowledge integration (estimate =−.01, 95% CI
[−.02, −.01]).
Hypothesis 4 posited that a focal team’s information brokerage
would weaken the negative relationship between inter-team competi-
tion and its inter-team knowledge integration. There was a significant
main effect of brokerage on inter-team knowledge integration (b=.55,
β=.53, p<.01). Furthermore, results showed a significant interaction
between inter-team competition and information network brokerage
(b=.12, β=.39, p<.01). A simple slopes test (Cohen et al., 2003)
indicated that the negative slope between inter-team competition
and inter-team knowledge integration was weaker when brokerage
was higher (+1SD;b=−.01, n.s.) as compared to when it was lower
(−1SD;b=−.06, p<.01). The difference between these slopes was
also significant (p<.01), thus supporting Hypothesis 4. This
interaction is illustrated in Figure 2 below.
Finally, in testing Hypothesis 5, we examined if the indirect
effect of inter-team competition on team innovation via knowledge
integration varied with the extent of the focal team’s brokerage. This
indirect effect was not significant when brokerage was higher,
+1SD; estimate =−.004, 95% confidence interval (CI) [−.01, .01],
as compared to when it was lower, −1SD; estimate =−.03, 95% CI
[−.03, −.02]. These effects provided support for our overall model.
Supplementary Analyses
First, we tested our hypotheses using alternative operationaliza-
tions of focal variables. For inter-team competition, these included
Table 1
Study 1 Means, Standard Deviation, and Bivariate Correlations
Variable MSD 1 2345678 9101112
1. Data site (inter-team network) size 10.43 7.70 —
2. Team size 9.40 2.20 .26*—
3. Project completion rate 3.12 0.88 −.07 −.20 —
4. PM gender
a
0.82 0.38 −.03 −.01 −.03 —
5. PM tenure 2.26 1.43 −.05 .09 .24*.16 —
6. Task interdependence 3.74 0.71 .23*.21 .13 .17 .22 (.66)
7. Task complexity 3.98 0.67 .28*.20 .11 .16 .32*.54** (.66)
8. Team identification 3.73 0.37 −.39*−.20 .38*−.16 .21 −.14 .02 (.75)
9. Competition in-degree centrality 6.67 3.91 .59** .38** −.02 −.08 .24 .08 .09 −.21*—
10. Information brokerage
b
−0.48 0.28 .79** .21 −.01 −.09 .14 .17 .17 −.32** .55** —
11. Inter-team knowledge integration 3.54 0.29 −.34*−.01 .14 .12 .11 .00 .08 .37** −.30*.21 (.80)
12. Team innovation 4.13 0.60 .08 .17 .22 .11 .17** .37** .40** .08 −.08 .08 .24*(.84)
Note. n =73. Reliabilities are provided in parentheses. PM =project manager.
a
0=female, 1 =male.
b
Brokerage is the negative of constraint.
*p<.05. ** p<.01.
4
At the team level, there was also a moderately strong raw correlation
between inter-team competition in-degree centrality and information
brokerage (r=.55), raising concerns of potential multicollinearity. To
understand the impact of this on our results, we took several steps and found
our results to be robust. Details are available in the additional online material
and appendix.
INTER-TEAM COMPETITION AND INNOVATION 35
unweighted degree centrality and degree centrality using recipro-
cated ties (where a competitive tie only exists if both teams report
each other); for brokerage, we ran models with effective size
(e.g., Soda et al., 2018). All results using these alternative measures
were consistent with our main analysis. Next, we tested if it might
be strength of network closure, where a team’s alters are densely
connected to each other (e.g., Ahuja, 2000), that may buffer the
adverse effects of competition. Analysis using a focal team’s
information ego network density confirmed that it was brokerage
and not closure that helped teams counter the knowledge deficit
caused by inter-team competition. Finally, we inspected the degree
of network overlap between a focal team’s competition and
information networks by dividing the number of overlapping ties by
the total number of information ties. The overlap median was 26.2%,
indicating that the two networks, while overlapping to a small extent,
are largely distinct. Analysis controlling for this overlap variable, or
subsample analysis after excluding a small number of teams with
high overlap did not change our results. Details of these analyses are
available in the additional online material and appendix.
Study 1: Discussion
Results from engineering teams showed that a high degree of
competition experienced by a focal team reduced innovation by
impeding its ability to acquire and utilize knowledge from the
network. However, occupying a brokerage position in the inter-team
information exchange network counteracted competition’s adverse
effects. Although Study 1 has strong external validity, it also has
important limitations. First, as a field study, it is unable to fully
address endogeneity concerns and, thus, is limited in terms of
providing causal inference. Second, it lacks insights into the
“microprocesses”of how a focal team utilized information brokerage
to combat the negative effect of inter-team competition (e.g., Grosser
et al., 2019). Furthermore, managerial ratings of team innovation
may not be entirely objective. To address these limitations, we
conducted Study 2, a team-based network experiment requiring
teams to develop innovative proposals, which in turn, were
objectively coded for innovation by trained and independent judges.
Table 2
Study 1 Results of Regression Analysis
Variable
Inter-team knowledge integration Team innovation
Unstandardized
coefficient b
Standardized
coefficient β
Unstandardized
coefficient b
Standardized
coefficient β
Control variables
Inter-team network size −.01 (.01)** −.51 (.20)** .01 (.01) .12 (.09)
Team size .01 (.01) .02 (.12) .04 (.03) .15 (.11)
Project team completion rate −.02 (.04) −.07 (.14) .10 (.06) .14 (.10)
Project manager gender
a
.14 (.11) .19 (.14) .02 (.10) .01 (.06)
Project manager tenure −.01 (.01) −.04 (.10) −.01 (.02) −.02 (.11)
Task interdependence −.02 (.04) −.04 (.10) .25 (.12)*.28 (.14)*
Task complexity .07 (.04) .16 (.10) .18 (.15) .20 (.17)
Team identification .16 (.04)** .20 (.04)** .05 (.15) .03 (.09)
Independent variable
Inter-team competition in-degree centrality −.03 (.01)** −.43 (.14)** −.03 (.01)** −.17 (.06)**
Moderator
Inter-team information brokerage/structural
hole
.55 (.20)** .53 (.14)**
Interaction terms
Competition in-degree Centrality ×
Information Brokerage
.12 (.04)** .39 (.13)**
Mediator
Inter-team knowledge integration .37 (.07)** .18 (.03)**
R
2
.36** .33**
Note. Both unstandardized and standardized coefficients are reported; standard errors in parentheses; n(individual) =689;n(teams) =73. Controlling
for the firm or data sites (using two and six dummy variables, respectively) did not alter any of these findings. For the sake of brevity, we have not
included them in this table.
a
Project manager gender, 0 =female, 1 =male.
*p<.05. ** p<.01.
Figure 2
Interaction Between the Degree of Inter-Team Competition and
Information Brokerage on Inter-Team Knowledge Integration
36 YAN, VENKATARAMANI, TANG, AND HIRST
Study 2: Method
Study Design
Study 2 was a team-based lab experiment conducted using an
online simulation. In each trial, groups of three teams comprising two
members each entered our experimental platform synchronously
to engage in a knowledge-intensive innovation simulation (these
three teams thus formed an inter-team network). In each inter-team
network, our focus was on one focal team (Team Red) that received
all our study manipulations and whose outcomes we were interested
in. We manipulated our independent variable—the degree of inter-
team competition faced by this focal team in the three-team
network—and our moderator, the focal team’s information brokerage
position. Teams then had the opportunity to interact with each other
to engage in the full spectrum of knowledge exchange activities, such
as making inquiries,sending and receiving knowledge, making deals
and even obstructing knowledge flow. Finally, using the knowledge
they were given and that they acquired during the simulation, each
team created an innovation proposal, which was coded by
independent judges to derive our dependent variable.
We created a 2 (high-competition condition: focal team facing
many competition ties vs. control condition: focal team faces zero
competition ties) ×2 (focal team as broker in the inter-team
information network vs. focal team as nonbroker) factorial design
using Empirica, a web-based multiplayer interactive experimental
platform (Almaatouq et al., 2021). For the main task, we adapted
materials from the Windy City Theatre paradigm for open-ended
tasks such as team creativity and innovation (Hoever et al., 2012,
2018;Parke et al., 2022). In this paradigm, teams are tasked to create
an innovative solution to improve business at a historic yet
struggling theatre. Each team received some common knowledge
(shared among the three teams) and some unique knowledge that
was only accessible to each team individually. While not necessary
to complete their proposals, it would be helpful if the teams
interacted with one another and shared/acquired different types of
knowledge from each other in developing their solutions. Overall,
this paradigm simulates interactions like those among real
organizational teams (such as teams in Study 1), wherein teams
tend to have both shared and unique knowledge and where informal
access to more diverse information provides distinct advantages in
tackling open-ended problems and creating innovation (Anderson et
al., 2014;van Knippenberg, 2017).
Data and Sample
972 individuals (51.5% female, 47% male, and 1.5% other) were
recruited from Prolific. Their average age was 36.8 years; 64.9%
were Caucasian, 15.1% were Asian/Pacific Islander, 7.3% were of
African descent, 2.6% were Hispanic or Latino, and 10.1%
identified as mixed or chose not to report. We recruited English-
proficient, working adults (average working experience of 12.5
years) who received $13 in compensation. Human subjects research
approval was obtained at the second author’s institution (Protocol
#1666940-2).
Task and Procedure
Each trial of the simulation required six participants to enter the
experimental platform synchronously. Participants were randomly
assigned to one of three 2-member teams denoted as Team
Red, Team Blue, and Team Green. Thus, the six participants
comprised an “inter-team network”of three teams, each with two
members. Our sample consists of 162 such inter-team networks,
with one of these teams (randomly chosen to be Team Red)
being the focus of all our manipulations and main analyses. It is
thus important to note here that each simulation run with six
participants in three teams yielded one data point (i.e., of Team
Red) in our sample.
Participants were informed that they were special project teams at
Riverside Theatre, a historic yet struggling theater in Chicago.
Participants read that due to declining interests in conventional
theatre, the managing director had tasked these three teams to
conduct independent research and propose innovative solutions to
improve the theatre. They were told that innovative solutions were
those that were novel and original, as well as useful and
implementable, and that based on their performance, teams could
earn a monetary reward. Participants read that each team had some
unique pieces of knowledge but that, to develop a comprehensive
solution, it would be helpful if teams gathered and utilized
knowledge from the other two teams as well.
To create their proposal, each team was presented with:
(a) general background knowledge about the theatre that was
available to all teams (e.g., theater layout, last year’s ticket sales),
and (b) four pieces of unique knowledge from their own research.
We conducted a pretest with 120 Prolific participants to ensure
that each team was given unique knowledge of equal value (please
see additional online material and appendix, p. 4). For instance,
Team Red had knowledge about the bars and restaurants near the
theatre and that the theatre could benefit from establishing local
partnership relationships with these hospitality venues; Team
Blue had knowledge about which type of shows had the highest
and lowest profit margins; Team Green had knowledge about a
local high school and senior care home that could potentially be
beneficial for the theatre’s outreach program. After reading the
information available to their respective teams, participants read
that the managing director had proposed a meeting among the
three teams. Teams then read about rules for the interaction phase,
where we inserted manipulations of inter-team competition
and information brokerage. Next, the three teams interacted for
about 14 min where they could discuss and negotiate to send
and/or receive information. After discussion, each team was
given 12 min to write their proposal. On average, the simulation
lasted 60 min.
Our theory focuses on how a focal team’s knowledge integration
and innovation are affected by the degree of competition it faces
from peer teams and how it is mitigated by being an information
broker between teams. Hence, out of the three teams in each trial, we
randomly chose Team Red to be the target of all our manipulations.
Thus, Team Red received one of the four manipulations based on
our 2 ×2 design.
Manipulation of Team Red’s Information Brokerage
We manipulated information brokerage using the communication
structure among the three teams in the between-team discussion
section (Brands & Mehra, 2019;Greenberg, 2021). In the focal team
as information broker condition, Team Red members read that the
managing director had chosen them to be the discussion coordinator
INTER-TEAM COMPETITION AND INNOVATION 37
among the three teams and was presented with a network diagram
that visualized their brokerage position. They read, “As this picture
illustrates, your team is the ‘bridge’that connects the other two
teams, Team Blue and Team Green. Only your team can
communicate directly with them, and they cannot communicate
with each other without going through your team.”Team Blue and
Team Green participants in this condition were shown the same
diagram but read that “As this picture illustrates, Team Red is
the ‘bridge’that connects your team with the other team. You can
directly communicate with Team Red but not with Team Green
[Blue]. In other words, if you want to communicate with Team
Green [Blue], you will have to do so indirectly via Team Red.”
Therefore, in the discussion phase of the simulation, Team Red
had two chat boxes to communicate with both Team Blue and
Team Green, whereas Team Blue and Team Green had only one chat
box—that is, to communicate with Team Red.
In the focal team as nonbroker condition, all three teams read
that the managing director had chosen an open discussion format.
They were shown a diagram of a fully connected communication
structure and read that, “As this picture illustrates, every team will be
connected with each other, and your team can freely communicate
with the other two teams.”Accordingly, every team had two chat
boxes to directly communicate with the other two teams.
Manipulation of Team Red’s Degree of Inter-Team
Competition
Consistent with our conceptualization of competition as occurring
dyadically between teams and our centrality-based (i.e., aggregated
number of incoming competition ties) operationalization in Study 1,
we manipulated the degree of competition faced by Team Red
by varying the number of between-team dyadic competitive ties
it had. Specifically, given that our inter-team networks in each
simulation consisted of three teams each, the highest and lowest
possible number of competitive ties Team Red could have were two
and zero, respectively. Thus, we maximized the separation between
the conditions such that Team Reds faced the maximum possible
competitive ties (i.e., two, with both Team Blue and Team Green)
in the high competition condition or the least possible competition
ties (zero; with none of the other two teams) in the no-competition
condition.
In designing our competition manipulation, we were guided by
prior work on the antecedents of dyadic competition. This research
suggests that between-team dyadic competition, where one team
strives to outperform another team (To et al., 2020), can be the result
of multiple factors. For example, it could be induced by dyadic
outcome interdependence (e.g., two teams competing for economic
incentives or scarce resources), other socio-relational factors such
as dyadic performance history (Kilduff et al., 2016), emotional
arousal specific to a pair of teams (Ku et al., 2005), or a combination
of these factors (Garcia et al., 2013,2019). Accordingly, in
providing an effective manipulation, we induced competitive ties by
providing Team Reds with information about their between-team
dyadic history with Teams Blue or Team Green as well as dyadically
structured incentives to outperform specific target teams. No such
information about the history or incentives were provided to teams
in the no-competition condition.
In the focal team facing high degree of inter-team competition
condition, Team Red faced the maximum degree of competition—that
is, two competitive ties, with both peer teams in their three-team
network. Team Red participants read that,
Your team, Team Red, has a competitive relationship with both Team
Blue and Team Green, separately. In the past, the Managing Director
has organized similar special projects, and your team has experienced
very intense competition with both Team Blue and Team Green. Hence,
in this special project, your team is very motivated to outperform each
of them.
They were informed that to outperform the other teams, they needed
to have a more innovative proposal than the other teams, as well as
gather more pieces of relevant information that would be helpful
for the proposal. Accordingly, they read that,
The proposal your team creates will be rated on innovativeness by the
researchers on a 7-point scale. In addition, for every piece of knowledge
your team acquires more than Team Blue and Team Green respectively,
your team will be awarded 3 points, which will be added to your team’s
final score, and will be used to determine your monetary bonus at the
end of the study.
Thus, if Team Red secured more overall points than Teams Blue
and/or Green, it could receive a higher bonus than each team,
respectively. Therefore, in line with prior research on between-team
dyadic competition, our competition manipulation provided
information about the past competitive history between Team
Red and the other teams individually and offered dyad-specific
economic incentives for Team Red to outperform the other two
teams separately.
5
In the focal team facing zero degree of inter-team competition
condition (i.e., the no-competition or control condition), all three
teams read that their teams had been set up anew by the managing
director and that their team was working hard to build up its reputation
and credibility. Participants were informed that they needed to have an
innovative proposal and that it would be helpful to acquire knowledge
from the other two teams in the discussions. Similar to the competition
condition, they were informed that their team’s proposal would be
rated on innovativeness by the researchers on a 7-point scale.
However, different from the competition condition, where they would
be awarded for every piece of information they collected more than
their competing teams, participants in the control or no-competition
condition were told that they would be awarded 2 points per piece of
knowledge they acquired, which would be added to the final score.
All teams would then have the opportunity to receive a monetary
bonus at the end of the simulation based on their overall score. Thus,
the no-competition condition had a similar setup but did not mention
or contain any language related to “outperforming”or competing
with the other teams. By having a similar point-based system
5
To ensure the strength of our manipulation, we conducted a
supplementary study with online participants from Prolific Academic
(n=120). Results showed that compared to only dyadic economic incentives
or only competitive history, our combination approach indeed created the
strongest dyadic competitive desires and striving to outperform a target
(details available in additional online material and the appendix, p. 5).
Therefore, we used this combination approach in our manipulation. We also
pretested the language to ensure that Team Red participants understood that
they had two between-team dyadic competitive ties with Team Blue and
Team Green, separately. Team Blue and Team Green participants in this
condition read similar instructions except that their team only had one
competitive relationship, i.e., with Team Red. We thank an anonymous
reviewer for this suggestion.
38 YAN, VENKATARAMANI, TANG, AND HIRST
and chance to receive a monetary bonus, we kept the competition
and control conditions broadly comparable to avoid any potential
confounds. Figure 3 shows the diagrams presented to participants
across the four conditions.
Inter-Team Discussion and Knowledge Exchange
After participants read both the manipulations, the three teams
entered a virtual chatroom interface for 14 mins where they used
dedicated chat boxes to communicate with another specific team
depending on the brokerage condition. Aside from the between-
team chat boxes, each team also had an additional chat box to allow
team members to communicate and strategize internally. On this
interface, participants also had access the knowledge available to
them, a diagram that illustrated their communication structure and
competitive relationships (as shown in Figure 3) and a spreadsheet
that allowed them to keep track of all the information. Figure 4
presents a snapshot of this interface.
Team Innovation Proposal
After the inter-team discussion, each team was provided
a collaborative text editor (similar to a Google Doc) for the
two members to create their final team proposal. To facilitate
their writing, they had a chat box where they could communi-
cate ideas.
Measures
Inter-Team Knowledge Integration
Across the 162 trials of the simulation, teams sent a total of 7,822
messages (M=48.28, SD =14.48). Two research assistants (RAs)
blind to the hypotheses coded the messages between teams to create
our mediator, the focal team’s inter-team knowledge integration. We
calculated the number of pieces of true knowledge acquired by each
team, with the logic being that when teams acquire a true piece of
knowledge (as opposed to partially true or an outright false one),
they are most likely to also utilize it.
6
Hence, RAs identified all the
messages through which the focal Team Red received a true piece
of knowledge from another team and calculated the total number of
pieces of such knowledge acquired. The RAs were given detailed
information about the different pieces of knowledge and first coded
a subset of the chat data together (n=30) to establish a consensual
coding standard. They then coded all remaining messages indepen-
dently and achieved very good interrater agreement (minimum R
wg
=
.90). Hence, we took the average score of the two RAs.
Figure 3
Study 2 Experimental Conditions (Focal Team Red)
Note. All manipulations are directed only at Team Red in all conditions. Red lines denote competitive ties and black lines denote communication ties. Each
icon represents a participant such that each team is comprised of two members. Top left: control/no-competition and nonbroker condition; top right: control/
no-competition and information broker condition; bottom left: high competition and nonbroker; bottom right: high competition and information broker. See
the online article for the color version of this figure.
6
We also analyzed our data with an alternative operationalization of
inter-team knowledge integration by adding independent RA coding of
teams’utilization of the knowledge acquired in their team’sproposals,and
team’s self-reported degree of inter-team knowledge integration. Results
are substantively the same, showing that the interaction applies robustly to
both the acquisition and utiliz