Outside in, inside out: The impact of knowledge heterogeneity, intra- and extra-organizational ties on innovative status

Abstract

The purpose of this paper is to explore how what people know and whom they know outside their organization shapes their innovative status among their colleagues. Combing innovation and network theory, we suggest that individuals that have diverse, heterogeneous knowledge relative to their colleagues and who can draw upon diverse contacts from across and outside their organization will have high status among their colleagues for developing new, creative solutions. The analysis is based on a comprehensive study of the knowledge and networks in an office of consulting engineers. Implications and lines of extensions are outlined for future research and theories on networks and innovation.

Full text

Paper to be presented at the Summer Conference 2009
on
CBS - Copenhagen Business School
Solbjerg Plads 3
DK2000 Frederiksberg
DENMARK, June 17 - 19, 2009
OUTSIDE IN, INSIDE OUT: THE IMPACT OF KNOWLEDGE HETEROGENEITY,
INTRA- AND EXTRA ORGANIZATIONAL TIES
Ammon Salter Salter
Imperial College Business School
a.salter@imperial.ac.uk
Paola Criscuolo
Imperial College Business School
p.criscuolo
Linus Dahlander
Stanford University
linusd@stanford.edu
Abstract:
The purpose of this paper is to explore how what people know and whom they know outside their organization
shapes their innovative status among their colleagues. Combing innovation and network theory, we suggest
that individuals that have diverse, heterogeneous knowledge relative to their colleagues and who can draw upon
diverse contacts from across and outside their organization will have high status among their colleagues for
developing new, creative solutions. The analysis is based on a comprehensive study of the knowledge and
networks in an office of consulting engineers. Implications and lines of extensions are outlined for future
research and theories on networks and innovation.
JEL - codes: M1, O3, O

1
O
UTSIDE
I
N
,
I
NSIDE
O
UT
:
T
HE
I
MPACT OF
K
NOWLEDGE
H
ETEROGENEITY
,
I
NTRA
-
AND
E
XTRA
O
RGANIZATIONAL
T
IES
ON
I
NNOVATIVE
S
TATUS
This version: 8 October 2008.
Word count: 9,133
PAOLA CRISCUOLO
Imperial College London
LINUS DAHLANDER
Imperial College London
Stanford University
AMMON SALTER
Imperial College London
Corresponding author:
Paola Criscuolo
Imperial College Business School
Imperial College London
Tanaka Building
South Kensington Campus
London, SW7 2AZ, UK.
Email: p.criscuolo@imperial.ac.uk
ABSTRACT
The purpose of this paper is to explore how what people know and whom they know outside their
organization shapes their innovative status among their colleagues. Combing innovation and
network theory, we suggest that individuals that have diverse, heterogeneous knowledge relative
to their colleagues and who can draw upon diverse contacts from across and outside their
organization will have high status among their colleagues for developing new, creative solutions.
The analysis is based on a comprehensive study of the knowledge and networks in an office of
consulting engineers. Implications and lines of extensions are outlined for future research and
theories on networks and innovation.

2
INTRODUCTION
Why are some individuals more likely than others to gain innovative status within an
organization? In this paper, we focus on a central facet in determining an individual’s innovative
status among their colleagues – their ability to be recognized by their colleagues for developing
new, innovative solutions. Research on this topic has tended to focus on how certain network
positions inside organizations positively influence individual outcomes (Burt, 1992), such as
promotion (Podolny and Baron, 1997), wages (Burt, 2004), and innovativeness (Obstfeld, 2005).
In much of this work, individuals are assumed to gain knowledge-based advantages over their
colleagues by virtue of being more connected (Freeman 1979), being situated on central
information flows or spanning structural holes (Burt 1992), yet there are few studies that
investigate what people know and how it differs from their colleagues. Indeed, scholars often
treat the social structure of network as synonymous with its knowledge structure (Tortoriello,
McEvily, and Krackhardt, 2008). Accordingly, what individuals know relative to others and to
whom they communicate with may offer very different network patterns. Given this fact, our
current understanding of networks and their influence on performance often fails to provide
information on the nature of the knowledge-based advantages that some individuals possess over
individuals working in the same organization. Because of a rich tradition of exploring the
consequences of certain positions in networks inside organizations and in one geographical
location (Burt, 2004), much research has also overlooked external ties that individuals draw upon
in their work. An internal focus may be sensible in stable, large organizations operating in mature
markets; but in sectors, where individuals work in teams that span many different parts of the
same organization or involve collaboration with external actors, this approach may provide a
partial picture of the building blocks of structural advantage that comes to individuals through the
possession of particular network resources.
In order to increase our understanding of the relationship between networks and
innovation, we build on two strands of literature that have rarely informed one another. The first

3
of these concerns the nature of the knowledge hold by individuals relative to their colleagues.
Research has shown that individuals who draw upon contacts with heterogeneous knowledge
have higher innovative performance than their colleagues who have less heterogeneous contacts
(Rodan and Galunic, 2004). This approach highlights the cognitive advantages that individuals
can gain from a heterogeneous set of contacts. In doing so, it picks upon the information benefits
that individuals’ gain from their network resources, which is a central element in the brokerage
theory. The second strand that has been given attention in past studies of innovation is the role of
intra- and extra-organizational network resources. Early work by Allen and others (Allen, 1977;
Allen and Cohen, 1969; Tushman, 1977) suggested that those individuals who can act as
gatekeepers between the internal and external parts of the organization are more likely to be
innovative. More recently, Cross & Cummings (2004) and Perry-Smith (2006) have demonstrated
that individuals possessing a high number of relationships across organizational departments have
higher innovativeness and creativity. To date, however, most studies count the raw number of
such ties, despite a growing consensus that the character of such external ties is vital.
These two approaches are valuable in themselves, but we argue that the full potential has
not been leveraged as they have rarely been integrated. Since the ability of an individual to be
innovative is likely to be both a function of what they know and whom they know, the lack of
integration between these perspectives leaves our theoretical understanding of the role of
networks on innovation incomplete. To advance theory, we combine these two approaches,
focusing on both the nature of knowledge that an individual possesses relative to their colleagues
and the diversity of their intra- and inter-organizational network resources. By combining these
two strands of extant literature, we are also able to explore how these factors interact to shape an
individual’s ability to gain innovative status.
Our analysis is based on a comprehensive network study of an Arup office, a world-leading
consulting engineering firm. The research site covers a group of individuals working on projects
that draw from resources inside and outside the organization. In the analysis, we combine data

4
from a network survey, the organizations’ expert location pages, and human resources records. In
order to test our hypotheses about innovative status, we use a negative binomial model, where an
individual’s knowledge heterogeneity, and the diversity of their intra- and extra-organizational
ties explain their innovative status among their colleagues. We find some statistical evidence for
our hypotheses.
KNOWLEDGE HETROGENEITY AND EXTERNAL TIES
The network literature has often equated social structure and knowledge – yet what people know
should not be taken as given by the nature of their social structure. The value of what an
individual knows has a relational character (Padgett and Ansell, 1993). Accordingly, individuals
who possess unique knowledge compared to the colleagues know something that is both rare and
potentially valuable to their organization (Blyler and Coff, 2003). In contrast, individuals that
have expertise in similar areas to their colleagues within the organization will gain less of an
advantage from that knowledge because it is likely that many other individuals within the
organization are able to cover or even substitute that individual in his or her work tasks.
1
Given
this fact, the benefits that individuals can extract from their knowledge has a strong relational
quality that can only be appreciated once an understanding of what an individual knows relative
to their colleagues is uncovered. To date, past research has only made limited progress in
exploring the relational character of individual’s knowledge and its impact on the status of
individuals inside their organizations.
Since knowledge is unequally and widely distributed amongst individuals, a central
challenge for organizations is to find new ways to access and productively combine this
knowledge (Hayek, 1948). As Schumpeter (1912/34) suggested, innovation is fundamentally a
combinatorial process that involves bringing together previously discrete or unconnected bodies
1
Of course, an individual whose knowledge is totally unique or very far removed from the organization’s
knowledge set may also find that their knowledge is not valued appropriately by their colleagues who have
little knowledge or interest in their area. An example would be a social constructivist in working in a
department of economics.

5
of knowledge. Such new combinations can help to unlock applications that were previously
unseen. In this context, the ability of individuals to discover new solutions to problems is related
to what they know and can do, and their ability to align their knowledge and expertise with the
information and resources they can muster from their social network. In searching for
innovations, individuals that can draw upon diverse or heterogeneous sources of ideas are more
likely to be able to bring together information from disparate bodies of knowledge into novel
outcomes (Burt, 2005; Obstfeld, 2005; Rodan and Galunic, 2004). As a result, the diversity of
knowledge that individuals can draw upon as well as the relative character of that knowledge with
respect to their work environment can help to shape the innovativeness of an individual within
their organization. Therefore, understanding the relational and heterogeneous character of an
individual’s knowledge and how this shapes their ability to gain status with their colleagues is a
key question that remains to be fully theorized.
An expanding body of literature investigates how individuals’ network positions within an
organization influence a range of outcomes (Borgatti and Foster, 2003; Hansen, 1999; Ibarra,
1993; Podolny and Baron, 1997). Because of its focus on the internal organization, much of this
literature treats the firm (often situated in one geographical location) as a sealed environment
without an external face. One theoretical concern with this approach is that by investigating
internal networks, we may stand short in explaining how some individuals draw upon network
resources from other parts of their organizations or from external organizations. As Brass et al.
(2004: 808) point out, “Actors who perform […] bridging roles are likely to know more and to
have influence in the larger, external network, but they may be peripheral (and expendable) to the
internal networks of the groups they belong to”. The boundary spanning literature also draws
attention to ties that span organizational boundaries which enable access to a diversity of ideas
and resources (Allen, 1977; Allen and Cohen, 1969; Tushman, 1977). We argue that the full
potential of this line of investigation has not been fully realized as earlier research has only
considered the raw number of external ties without considering the character of such ties. Indeed,

6
almost forty years ago, (Allen and Cohen, 1969: 18) argued that “the content processed by the
various gatekeepers in research and development laboratories should be examined in more
detail”. To advance this literature, we distinguish between (1) intra-organizational and (2) inter-
organizational ties.
Intra-organizational ties are important for innovation because they allow organizations to
combine knowledge from different organizational sub-units. This is echoed in Kogut and
Zander’s (1992) contention that a firm can be understood as a social community specializing in
the creation and transfer of knowledge.
Past research shows that units that are able to draw upon
rich relationships with other divisions have higher innovative performance (Tsai, 2001). The
same advantage may hold for individuals. In Cross and Cummings’ (2004) study, individuals
with high number of ties to other departments had higher innovative performance than individuals
with inward-looking networks. Such inter-organizational ties can provide individuals with access
to pools of knowledge within the organization, chances to combine knowledge from different
areas, and opportunities to gain legitimacy for their activities among their colleagues. All of these
factors will help to ensure that such ties are an important factor in explaining the innovative status
of an individual among their colleagues.
Inter-organizational ties have also been shown to play a key role in shaping the
innovativeness of individuals and organizations (Ahuja, 2000; Baum, Calabrese, and Silverman,
2000; Powell, Koput, and Smith-Doerr, 1996). Recent efforts by firms to ‘open up’ their
innovation activities are based on the idea that individuals should make extensive efforts to
network externally to capture and integrate knowledge from outside the firm (Chesbrough, 2003).
Firm-level studies show that there are innovative benefits of drawing upon a broad range of
external sources of innovation, compared to relying solely on internal or a small number of
external sources (Laursen and Salter, 2006). Within firms, however, there is likely to be a broad
range of external ties that are held by individuals and the nature of these ties may influence an
individual’s ability to perform in their assigned role. Such ties can also act as a powerful stimulus

7
to help support internal efforts to innovate by offering resources and expertise that are not
available within the organization. Inter-organizational ties can be developed through a diverse set
of external relationships, including working on projects; through having worked together in the
past; shared educational experience; joint membership in clubs and societies; or through their
involvement in communities of practices that span beyond the boundaries of the firm. In addition,
the very nature of the workplace may also have changed during the last decade (Evans, Barley,
and Kunda, 2004). For instance, in Allen’s seminal study of boundary spanners, engineering
networks were considered to be largely local in nature, restricted to relationships within a single
organization or division. In his study, many individuals worked in teams that spent long periods
working together in the same team (in one case 20 years). With the rise of more fluid, fast-paced
and mobile work environments, individual’s networks are likely to be highly diverse, permeable
and external-oriented rather than monolithic, closed and internally focused (Saxenian, 1994). At
times, external ties may offer richer and more diverse sources of ideas and support than the
resources that can be gained from drawing on knowledge from the internal network. Individuals
who are able to draw upon a diverse set of external ties may be able to call on network resources
that their more inward looking colleagues are unable to. Thus, these external ties may be a ticket
to acquire innovative status within an organization.
THEORY AND HYPOTHESIS DEVELOPMENT
Inter-organizational networks
The ability of a firm to create new innovations often hinges upon flows of knowledge facilitated
through person-to-person relationships (Szulanski, 1996). Attempts to explore intra-
organizational networks have shown that units’ positions within the firm and the network
resources they can draw upon from other units, profoundly shape their ability to innovate and
share knowledge (Hansen, 1999; Monteiro, Arvidsson, and Birkinshaw, 2007; Tsai, 2001). These
works show that organizations drawing upon resources from other divisions often requires
complex deal making. Such arrangements create expectations of future reciprocity and require

8
investment on the part of the unit to capture the benefits of the network ties that are formed (Tsai,
2002). To be sure, building links to actors in other divisions of these organizations can be a
daunting task for individuals. There are often significant barriers between organizational units;
colleagues in other divisions may be socially and culturally distant, located in departments with
different management and resource systems. Gaining access to resources or individuals from
other departments may require effort to build norms of reciprocity with colleagues who they do
not see face-to-face and/or who have different reporting or incentive structures. Such barriers are
likely to heightened in professional services firms where advice is often strategically received and
given (Lazega, 2001).
The challenge of building and sustaining inter-organizational ties means that such network
resources are likely to be unevenly distributed among individuals within an office (Perry-Smith,
2006). However, as Cross and Cummings (2004: 929) suggest “in complex work that demands
integration of specialized knowledge, people with ties crossing both organizational and
departmental boundaries are likely to find more relevant information and be more effective in
solving problems”. In order to reap these advantages, some individuals may invest in building
rich and deep networks across their organization, whereas others rely on only a small number of
key colleagues in a single division. Rather than investigating the sheer number of intra-
organizational ties, we argue that it is important to explore the organizational affiliation of these
ties (Geletkanycz and Hambrick, 1997). Accessing knowledge from a diverse range of
departments can itself provide a different set of benefits from the possession of a large number of
intra-organizational ties. Holding relationships to individuals that span different units of the firm
allows individuals access to resources and information, rather than relying on people working in a
single external unit. As yet, the full implications of possessing diverse intra-organizational ties for
an individual’s innovative status has not been fully investigated and theorized.
There are three important and related arguments to support the idea that holding diverse
inter-organizational ties will play an important role in determining an individual’s innovative

9
status with their colleagues. The first advantage of inter-organizational ties is related to
knowledge re-use (Majchrzak, Cooper, and Neece, 2004). Individuals that have ties to individuals
in different parts of the organization allows an individual to tap into the broad range of skills and
capabilities available, enriching the pool of experience and knowledge that an individual can
draw upon to solve problems (Tsai, 2001). In contrast, an individual who has intra-organizational
ties concentrated in a few divisions would find the range of experience and knowledge available
to them to be highly constrained, leading them to search locally for solutions. In this context,
local search involves looking to close confidents or familiar bodies of knowledge rather than
exploring areas or bodies of knowledge that are more distant and involve combinations
comprising far-flung knowledge elements (March, 1991; Stuart and Podolny, 1996). As a
consequence of this local search, they may miss opportunities to re-use knowledge and
experience that has already been developed within the organization (Dearborn and Simon, 1958).
In turn, this could lead them to ‘re-invent the wheel’ on their projects (Veshosky, 1998). In
contrast, those individuals who can draw upon ideas and support from colleagues working in
multiple divisions are likely to be able to see opportunities to re-use the organization’s existing
knowledge in new circumstances.
Second, working with colleagues in different departments can expand the potential that
individuals will find new combinations by weaving together more diverse sets of knowledge than
can be created within the confines of a single office. Different technological domains within an
organization normally embody different content, norms and languages (Tushman and Scanlan,
1981). Only having ties inside the same functional area or department often results in inefficient
solutions to novel problems (Reagans and McEvily, 2003). Moreover, in many organizations,
capabilities are highly diffused across divisions. Accessing these capabilities will usually require
extensive interaction with individuals from these departments. Therefore, diverse intra-
organizational relationships can give individuals a ‘combinatorial’ advantage over their
colleagues with less well-developed and narrower intra-organizational networks.

10
Third, drawing upon resources from other parts of the organization can help facilitate
coordination and mutual understanding for an individual’s actions within the organization.
Indeed, research has demonstrated the importance of ties crossing departmental boundaries for
effective knowledge transfer within organizations (Hansen, 2000; Tsai, 2001). This can be clearly
seen in Ancona and Caldwell’s (1992) study of product development teams, where one of the key
factors for the successful delivery of an innovative project is the ability of the project leader to
win support for their project from other parts of the organization. Thompson (1967) also
suggested that under conditions of reciprocal interdependence, as in innovative projects,
coordination needs to occur between functions and hierarchical lines. Individuals that facilitate
such mutual adjustment may garner legitimacy and a higher profile for their work, allowing them
to call upon support and resources from other parts of the organization, as well as protecting their
efforts from internal roadblocks and external interference. This is likely to be especially important
in the case of innovative activities that may involve actions that are removed from the traditional
routines or ways of working in the organization. The combined effect of these advantages leads
us to posit:
Hypothesis 1: The greater the diversity of an individual’s inter-organizational
ties, the higher their innovative status among their colleagues.
Inter-organizational networks
The literature on inter-organizational networks focuses on how organizations access external
ideas and resources through formal organization-to-organization ties (Brass et al., 2004). Yet
these formal inter-organizational ties only cover a small part of the total range of an
organization’s external ties. Indeed, as Powell et al. (1996) indicates, when introducing the
attitudes to formal agreements for a CEO of a biotech firm, formal inter-organizational ties
represent "the tip of the iceberg - it excludes dozens of handshake deals and informal
collaborations, as well as probably hundreds of collaborations by our company's scientists with
colleagues elsewhere". Inter-organizational ties exist at many levels in the organization that may

11
or may not appear in their formally reported networks of affiliation (Ahuja, 2000). Moreover, it is
often difficult for organizations to map their employees’ external relationships, as individuals
may be reluctant to share this information with colleagues or because such relationships are
comingled with outside work interests.
Two related arguments support the notion that holding diverse external ties will influence
an individual’s innovative status. First, forming external ties with individuals from other
organizations increases the ability of the individual to get access to information about market and
technological developments (Tushman, 1977). Individuals with a rich and varied external
network get access to views, mindsets, perspectives and ideas that may diverge from accepted
norms within their focal organization. Such individuals can be important filters for an
organization to select, interpret and transmit information originating in the environment (Aldrich
and Herker, 1977). Accessing external contacts also expands the range of potential solutions to a
problem by providing access to skills and knowledge that are not held within the organization,
expanding the pool of resources available to the individual to solve problems and find new
solutions.
Second, experts often have a preference to be connected to individuals in external
communities over individuals in their local community (Marcson, 1960; Van Maanen and Barley,
1984). In a work place of specialists, there may thus be a cognitive bias in favour of individuals
that are different. For instance, Menon and Pfeffer (2003) argue that there is a preference for
outsiders where individuals over-value and over-use external knowledge, compared to rich
internal knowledge from which value can be captured much more easily. They argue that the
preference for external knowledge is the result of managerial attitudes towards the scarcity of
knowledge. Internal knowledge is more readily available and subject to greater scrutiny, while
external knowledge is scarcer, which makes it appear more special and unique. Given these two
arguments, we posit:

12
Hypothesis 2: The greater the diversity of inter-organizational ties of an individual, the
higher their innovative status among their colleagues.
Knowledge heterogeneity
A complementary view as to why some individuals gain innovative status suggests that a diverse
background provides a stronger foundation for learning. Research shows that individuals with
diverse interests are likely to have greater chance of being able to recombine the knowledge into
novel outcomes (Rodan and Galunic, 2004). The benefits of knowledge heterogeneity for
innovation is picked upon in Fleming and Sorenson’s (2004) notion of combinatorial search,
where they suggest that search processes that are more diverse have a greater likelihood for
yielding breakthroughs than search processes that move along single and established domains of
knowledge. Indeed, individuals whose interests and expertise span different areas have the
potential to see ways of transferring knowledge developed in one area to another. In Hargadon
and Sutton’s (1997) study of IDEO, they demonstrate the importance of bringing together
individuals with diverse knowledge to solve complex problems. It was only through brokering of
these ideas from one place to another that it was possible to make the leap of imagination to
realise the new solution. This finding is consistent with Burt’s arguments about the information
advantages of brokers. By connecting the disconnected, brokers gain access to unique information
and take ideas from places where they are common to areas they are new in. As Burt (2005)
suggests, opportunities for importing and exporting ideas across a network are often missed by
those whose experience is limited to single domains. In this sense, knowledge heterogeneity can
provide a cognitive advantage.
Knowledge heterogeneity also provides a social advantage. Individuals with high
knowledge heterogeneity are likely to be few in number within any organization, as keeping
aware of developments in several knowledge domains is difficult and specialization is often the
consequence of professional training. In other words, most individuals within a professional
organization are likely to be hedgehogs (individuals specialising in a few areas of knowledge)

13
rather than foxes (individuals whose knowledge spans many domains). However, for those
individuals who are foxes, they are likely to be critical gatekeepers of new sources of information
for their colleagues. Their breadth of interests or knowledge is liable to make them attractive to
their colleagues as they can offer insights and perspectives about potential solutions that are
somewhat distant from established norms. By having a diverse set of interests, these individuals
can also act as a map for others through a complex landscape of potential solutions, helping their
colleagues jump across a terrain of potential solutions to find ones that were outside their frame
of reference or experience. In this respect, the foxes can help speed up the search efforts of the
hedgehogs and therefore they are likely to gain high innovative status among their colleagues.
Thus,
Hypothesis 3: The greater the knowledge heterogeneity of an individual, the
higher their innovative status among their colleagues.
Knowledge heterogeneity and external ties
Who benefits the most from external ties? In Rodan and Galunic’s (2004) study, individuals that
have access to an heterogeneous knowledge base through their social network have higher levels
of innovativeness. This work suggests that it is not only the social structure that matters, but the
knowledge an individual can reach through their network that shapes innovativeness. In this
approach, heterogeneity stems from alters that an individual is connected to and not what the
individual knows relative to their colleagues. Here we attempt to extend this logic to explore the
contingent effect that knowledge may have on the individual’s ability to benefit from their
external ties.
Research on individual learning has shown that an individual’s prior knowledge and
experiences will influence the future search for new ideas (March and Simon, 1958). Even in
situations when there are benefits from interacting with external actors, it is difficult to search
more broadly outside the focal unit or department. We therefore argue that knowledge
heterogeneity has a contingent effect on an individual’s ability to make use of a diversity of

14
external ties. Because a diversity of external ties typically brings experiences from diverse
organizational settings with different languages, norms and practices (Geletkanycz and Hambrick,
1997), individuals with less knowledge heterogeneity may have difficulty connecting this
information to internal activities. This is accentuated in Cohen and Levinthal’s argument that
knowledge diversity “facilitates the innovative process by enabling the individual to make novel
associations and linkages” (Cohen and Levinthal, 1990). An individual with high knowledge
heterogeneity has a greater ability to understand opportunities for recombinations and by
extension to “lay claim on knowledge from networks” to use Kogut and Zander’s (1992)
language. We would therefore expect individuals with greater knowledge heterogeneity to benefit
most from external ties:
Hypothesis 4a: Knowledge heterogeneity moderates the effect of the diversity
of intra-organizational ties in such fashion that individuals with high
knowledge heterogeneity benefit most from intra-organizational ties.
Hypothesis 4b: Knowledge heterogeneity moderates the effect of inter-
organizational ties in such fashion that individuals with high knowledge
diversity benefit most from the extra-organizational ties.
METHODS
Research site
The importance of knowledge heterogeneity and external ties for innovation is likely to be
especially pertinent in professional service firms. Professional service firms work on the basis of
projects, often involving considerable interaction with clients (Maister, 1993; McKenna and
Maister, 2002). These projects require the skilled application of expert knowledge to unique or
small numbers of situations, usually in collaboration with other professionals and organizations
working together towards specific and time limited goal. New practices or innovation emerge in
this environment through collaboration among professionals on projects and new practices likely
to be shared with other professionals outside the firm through professional associations and
exchanges (Davies and Hobday, 2005; Dodgson, Gann, and Salter, 2007; Dougherty, 2004). In

15
order to deliver their projects, professional service firms rely on the abilities of their expert staff,
whose knowledge and capabilities have been developed through long-periods of formal
education, and on-the-job mentoring through joint working and problem solving with more senior
staff (Teece, 2003). For individuals working in these organizations, building up a reputation
among their colleagues for creative and inventive problem solving and innovation is critical for
career progression (Lazega, 2001). Professionals need to ensure they are recognised by their
colleagues for unique and specialised knowledge, and for their ability to apply this knowledge to
the challenges posed by potential clients. In professional environments, such reputation can
enable individuals to gain the authority and status necessary to capture organizational resources
and earn high rents from their expertise. Indeed, Lazega (2001) describes the efforts of
individuals to build up reputation on professional service firms as a ‘status competition’, whereby
professionals behave strategically to build high standing among colleagues. As such, professional
service firms provide a perfect context to explore the impact of knowledge heterogeneity and
external networks on innovative status. This is because these organizations involve almost totally
knowledge-based work and rely on collegiality between highly skilled individuals with
considerable autonomy, who are seeking status among their colleagues for personal and material
gain (Hitt, Bierman, Shimizu, and Kochhar, 2001).
This study is based on information on the knowledge and networks of a group of consultant
engineers working in a multidisciplinary office. The office is a subsidiary of Arup, a large
multinational engineering consulting organization with units in well over 50 countries. The site of
our study was formed in late 1980s and, although located in the home country of the firm’s
headquarters, it is geographically distinct and has a high degree of operational autonomy. Since
its foundation in the 1980s, the office has grown organically to over 250 people at the time of our
study. Like many professional service firms, Arup is a highly decentralized organization. For
example, the regional office bids for and manages its own set of projects; and is responsible for

16
recruitment and retention of employees. The office also has several internal units within it which
focus on different areas of engineering practice.
The context for the study was an attempt by senior staff to re-locate the office to a new
building. As the office has grown, its staff were spread out across three buildings and there was a
sense that there was poor communication between different locations. Our study was designed to
provide some information to the senior management team about future office layouts. In addition,
the senior team at the office was concerned about the position of the office in the wider intra-
organizational network. Being somewhat distant from the corporate headquarters, they were
unsure to what extent their staff were able to forge ties with other parts of the organization. Our
research shows the relationship with the office and Arup in general has been a close one,
involving several rounds of interaction before and after our study, and engagement in several
related research projects.
To address our questions of interest in this paper, we combined data from three different
sources: (1) a comprehensive network survey, (2) internal expert pages and (3) human resource
records. Using diverse sources of data allows us to avoid common-method bias that can “arise
from a common rater, a common measurement context, a common item context or from the
characteristics of the items themselves” (Podsakoff, MacKenzie, Lee, and Podsakoff, 2003 886).
An added advantage of using the company’s internal information is that it minimizes the risk of
only relying on self-reported data that could be biased due to social desirability.
Network survey. Drawing on the literature on social networks and several interviews with
Arup staff, we developed a comprehensive network questionnaire. The survey covered internal,
intra- and inter-organizational networks of individuals, picking up on the interests of our
industrial research partner. The first draft of the questionnaire was piloted with a focus group
containing 12 engineers from different units within the office. The survey was significantly
adapted afterwards to reflect the tone, language and pace of work in the office. In the first section
of the survey, we asked questions about three different social networks within the office. Three

17
name generators were used to elicit the names of people who were instrumental for (1) problem
solving activities, (2) searching for new ideas, and (3) helping in the implementation of a
solution. The exact wording of the name generator questions are reported in Table 1.
2
Respondents could select up to 15 names for each of these three networks from a drop down
menu listing all employees in the office. For all our network questions, we asked respondents
about the nature of their ties with respect to communication frequency, how often they worked
together in past projects, and whether they were friends. In this paper and in order to make the
analysis tractable, we focus only upon the search network, as this internal network is the one most
directly related to innovation.
In the following section of the survey, we used the same name generator questions to
solicit names of people working for Arup but not in the same office. Respondents could list up to
15 names for each of the three intra-organizational networks, therefore a maximum of 45
different contacts within Arup. For each individual, we asked them to write their name,
department and location in the firm, using a drop down list of departments and locations provided
by Arup. We then matched this information to human resources records to obtain accurate
information about the department of the individual whom they were in contact with, allowing us
to avoid errors on the part of respondents with respect to departments and/or names. The survey
also asked respondents to write the name of up to five external contacts who acted as a critical
source of knowledge (see last item in Table 1). Our pretesting of the survey suggested that few
individuals would list more than five people. For each individual mentioned, we asked
respondents to provide information on their affiliation. We classified the different external
organization in one of the following types: other consultant organizations; universities or public
research labs; contractors; suppliers of equipment or of software; public sector organizations;
2
We measured the correlation between the three different networks (problem solving, search and
implementation) to assess whether they capture the exchange of different resources. The QAP correlation
between the search network and the problem solving and implementation network were both significant
and equal to 55% and 57% respectively, which suggests that these networks track different knowledge
flows.

18
architects; and client organizations (e.g. private developers, housing associations). This list of
types of external organizations is developed through interviews with Arup staff. Extensive web
search for names of the external contacts mentioned by the respondents and his/her affiliation was
carried to obtain accurate information and to assess the nature of the business activity of the
external organization.
-------------------------------------------------
INSERT TABLE 1 ABOUT HERE
-------------------------------------------------
The survey was sent out to all 231 engineers and administrative staff in the office. We
used an electronic survey tool called Lime Survey to capture responses. The survey received 204
responses corresponding to an 88% response rate. It took between 30 minutes and one hour to
complete. In addition, after the first round of the survey, we checked to ensure that the most
central people in the network completed the survey. For central non-responders, we telephoned
them directly, helping to raise the comprehensiveness of the sample. To check for non-response
bias, we compared non-respondents to respondents and found no significant differences by grade,
tenure or group of the office.
Internal expert pages. We retrieved data from the firm’s expert pages to derive a measure
of knowledge heterogeneity. Arup’s employees are encouraged to provide and update a
description of their expertise on their personal profile appearing on the company’s intranet. An
example of one of the skills description reads:
My principal interests are the Structural Design of buildings, typically offices, leisure and
retail and residential including: medium rise structures up to 25 storeys; reuse and
conversion of existing structures, basements for accommodation and car parking; blast
walls for protection against terrorist threat. Other more specialist interests include
swimming pools and Aircraft Access docking for Boeing 747, 737 and Tornado jets.
As the example above indicates, these skills descriptions are quite rich in information about
knowledge of the individual. The information contained in the expert location system is

19
searchable and frequently used by Arup staff. Although this information is not formally validated
within the organization, there are strong incentives upon individuals to make sure this information
is accurate. Indeed, they will be expected to answer questions from their colleagues on the areas
they mention and failure to respond to such queries could lead to a loss of status with colleagues.
Moreover, the descriptions are checked annually during appraisals.
Unfortunately, for some of the respondents to the survey, we do not have information on
their skills description. Thus we lose 73 observations when matching this information to the
network survey. We checked whether this introduces bias in our analysis by testing whether the
distribution of the main individual’s characteristics, such as gender, grade, tenure, educational
background, and employment status, differs between the sample of respondents reporting
information on their skills and the sample of respondents who have not provided this information.
While there was no significant difference between these two groups in terms of gender, grade,
educational background, and employment status, we found a statistically significant difference in
terms of tenure. The average tenure of those individuals for which we have a description of their
skills was seven years compared to 1.4 years for those for which we do not have this information.
However, the correlation between tenure and knowledge heterogeneity and both intra- and inter-
organizational ties is not very high (maximum .4). This suggests that although these two samples
of respondents do differ in terms of their tenure in the organization, they do not necessarily differ
in terms of the three independent variables of interest in this study.
Human resource records. We also use the company’s human resource records to retrieve
information on grade, tenure, educational background, employment status (fulltime versus part
time), and membership in professional organizations, such as the Institute of Civil Engineering.
Measures
Dependent variable. We measure innovative status by calculating the in-degree centrality for
each individual, i.e. the number of people who selected that particular individual in the internal
search network. This network measure is also called ‘degree prestige’ (Wasserman and Faust,

20
1994) and it has been used in several previous studies to capture the status on an individual in a
network (Brass and Burkhardt, 1992; Lazega, 2000). Another indicator that is often used to
measure status is the one developed by Bonacich (1987), which accounts for the centrality of the
alters who nominate the focal individual. This measure, however, cannot be applied in directed
networks as the one we analyse in this study (Bonacich and Lloyd, 2001). Thus, our measure of
individual status attempts to capture the value that colleagues place on an individual for their
ability to provide new engineering solutions and novel ideas. This measure of innovative status
has the advantage of not being self-reported and nor is it dependent on the assessment of one or a
few managers (Ibarra, 1993; Obstfeld, 2005; Rodan and Galunic, 2004). Instead, it relies on the
judgement of all the other colleagues working in the same office.
Independent variables
Diversity of intra-organizational ties was derived by computing the Blau’s (1977) index (1-p
i
2
)
where p
i
is the proportion of contacts working in department i. The Blau’s index is designed to
handle categorical data, so we do not need to make any assumptions about how different Arup’s
departments are. Respondents mentioned a total 493 different individuals, working in 42 different
Arup departments. On average, individuals mentioned contacts working in two different
departments, with a maximum of 18 out of the possible 64 departments that exist within Arup.
3
Some individuals in our sample have a very diverse intra-organizational network; more than 30%
of individuals have connections in three or more departments.
Similarly, we calculate the diversity of inter-organizational ties by calculating the Blau’s
index when p
i
represents the fraction of external contacts working for one of the following types
of organizations: other consultant organizations; universities or public research labs; contractors;
suppliers of equipment or software; public sector organizations; architects; and client
organizations. A total of 130 different organizations and 233 unique individuals were listed by the
3
We obtained consistent results when we characterise the diversity of intra-organizational ties using
information on the geographical location of Arup colleagues mentioned by the respondent.

21
total sample. The majority of external organizations are contractors (40%), followed by university
or public research institutes (20%) and public sector organizations (20%). On average,
respondents mentioned 1.3 different types of organizations, with a maximum of 5.
We believe our two measures of diversity expand prior research, which has to date only
focused on the sheer number of intra-organizational contacts (Cross and Cummings, 2004; Perry-
Smith, 2006). However by simply counting the number of ties, one does not account for the
extent to which these ties provide access to a range of perspectives and organizations, which, as
shown by the figures above, can be very significant.
We derived a measure of an individual’s knowledge heterogeneity by exploiting the
information on the internal expert pages. We followed the approach suggested by Criscuolo et al.
(2007) and extract a list of keywords from the skills description of all the members of staff. From
this list, we selected the keywords that appear more than 10 times and derive a person by the
keyword asymmetric matrix (X) in which cell x
ij
> 0 if the ith person mentions keyword jth in
his/her skill description, and x
ij
= 0 otherwise. This matrix is then transformed into a symmetric
matrix whose ijth cell contains the number of keywords that both person i and person j have in
common in their skill descriptions. This symmetric matrix can be represented by a network,
where each node is an individual and two individuals are connected if they share at least one
keyword in their skills description. We measure an individual’s degree of knowledge
heterogeneity by computing his/her betweenness centrality in this network (Freeman, 1979). If g
ij
is defined as the number of geodesic paths between individual i and j, and g
ikj
is the number of
these geodesics paths that pass through k, k’s betweenness centrality is equal to:

≠≠
ij ij
ikj
kji
g
g with
Betweenness centrality is a measure of how often a node is located on the shortest path
between other nodes in the network. If an actor is between two other actors then it implies that
there is not a connection between the alters on the path connecting them. An individual with high

22
betweenness centrality shares keywords with many individuals who themselves do not share
common skills, which suggests that the focal individual has a very diverse combination of skills.
For example, the skills’ description of the individual with the highest betweenness centrality in
our sample contains reference to expertise related to the area of geotechnical and geotechnics,
tunnelling, urban regeneration, land reclamation and regeneration, development groundwater,
underground fires, building contamination, re-use and recycling of demolition materials. This
particular engineer works in the geotechnics division, but has also applied their skills to many of
the other engineering practices such as fire, environmental, transportation, water and urban
planning.
By using the information stored in the internal expert pages, we are able to extend
measures of knowledge heterogeneity, locating an individual’s knowledge in the context of their
work environment. In past network studies, the knowledge of an individual is often inferred from
their social network position. It is often assumed that individuals spanning structural holes have
heterogeneous knowledge because they are connected to individuals whom themselves are not
connected (Burt, 1992). In an important extension of this approach, Rodan and Galunic (2004)
examined the knowledge heterogeneity of key contacts. Although an important advance in itself,
this approach has the limitation of relying on individuals to infer the knowledge of their contacts
and it also relies solely on information from the network survey. In contrast, our measure offers a
more relational measure of knowledge heterogeneity because it locates the individual within the
knowledge base of the entire office. It also relies on secondary data, drawn from sources beyond
the network survey. Therefore, it may provide a more direct measure of the knowledge advantage
that the individual may posses relative to their colleagues than has been available in past studies.
Control Variables. We control for a number of individual characteristics that could affect
the baseline level of an individual’s innovative status. In particular, hierarchical position may
have a positive effect on the likelihood of achieving a higher status in terms of innovation. In
professional service firms, senior members of staff are likely to be highly skilled problem solvers.

23
Senior employees might be responsible for the work of many junior employees, which will give
them an insight into a wide range of projects that are not available to more junior colleagues. We
measure seniority on a four point-scale based directly on the company’s human resources dataset:
0 for junior staff; 1 for experienced staff; 2 for senior staff; and 3 for directors (including
associate directors). This variable is dummy-coded and the reference category is directors. We
also control for the individual’s tenure in the organization by including a variable equal to the
number of years since he/she joined the firm. Those individuals with long tenure in organizations
are often able to learn about a wide range of activities that are undertaken inside the firm. They
may have worked in several different departments, and on a wide range of projects. This
organizational experience should allow them to provide novel solutions to colleagues, which
other individuals with shorter experience in the organization would be unable to supply (Rollag,
2004). Similarly, individuals with a full-time position might be able to develop a greater and
deeper organizational experience than those employed on a part-time basis. We therefore include
a dummy variable in the regression, which is equal to 1 if the individual is a full-time employee
(Full time).
Another variable that could affect the ability of an individual to achieve high innovative
status is education. Individuals with a post-graduate education might have acquired a specialized
set of skills that enable them to help others in exploring new engineering solution and formulating
novel ideas. Differences in education were accounted for by using a three-category ordinal scale:
0 for a degree or diploma (Degree); 1 for a master degree (Master); and 2 for a PhD (PhD). This
variable is dummy-coded and the reference category is members of staff with a degree or a
diploma.
Being a member of a professional organization, such as the Institute of Civil Engineering,
could also provide individuals with technical knowledge and up to date information on
regulations that could translate into higher innovative status. We control for this by including a
dummy variable (Member of Professional Organization) equal to 1 if an individual is a member

24
of at least one professional organization. Engineering consulting firms is still largely a male
dominated environment. In this context, women might be perceived to be less acknowledgeable
by their male colleagues. To control for this gender effect, we include a dummy (Gender) equal to
1 if the individual is a male. The number of characters in the skills description (Skills Description
Length) was also included as a control because an individual’s knowledge heterogeneity score
may be shaped by the amount of text they write in their skill description. Finally, all models have
been estimated with 14 division dummies that capture differences in size and engineering
knowledge among the different engineering practices within the office.
RESULTS
Table 2 reports the descriptive statistics for the variables used in the models. The average in-
degree centrality in the search network is 5.2, i.e. individuals were chosen on average by over five
colleagues for their ability to provide innovative solutions. On average, members of staff have
spent seven years in the organization, which reflects the recent growth of the office. Very few
engineers have a PhD, but a larger share obtained a Masters degree. Moreover, more than one
third are members of a professional organization. As mentioned earlier, women represent a small
fraction of employees in this organization. As expected, tenure and seniority in the organization is
positively and significantly associated with innovative status. Knowledge heterogeneity and
diversity in both inter and intra-organizational ties are also positively associated with innovative
performance at five percent level of significance. It is interesting to note these independent
variables are also positively correlated with seniority and tenure in the organization, but that these
correlations are not very high. Since most correlation coefficients are below 50 percent, we do not
need to be concerned about multicollinearity problems in the regressions.
-------------------------------------------------
INSERT TABLE 2 ABOUT HERE
-------------------------------------------------

25
Since individual innovative status – our dependent variable – was measured with in-degree
centrality, we tested our hypotheses using a count data model, which is appropriate when the
dependent variable is in the form of an event count. Table 3 shows the coefficient estimates
derived using a negative binomial model. Although the mean of our dependent variable is of
similar magnitude to its standard deviation, the negative binomial model seems to fit the data
better than the Poisson model.
4
To account for the fact that individuals of similar grade in the
organization might show a similar probability of achieving higher status, we estimated negative
binomial models with clustered errors for the different grade classes (Moulton, 1990).
5
Model 1 reports the results of the negative binomial model including only the control
variables. Some of the controls merit attention. Job grade is highly significant suggesting that
being a junior member of staff decreases the expected number of nominations by other colleagues
in the search network by five, holding all the other variables constant at their means. Employees
on full-time contracts and those holding a PhD are significantly more likely to achieve higher
innovative status. However, we find no significant negative effects for the innovation status for
women.
-------------------------------------------------
INSERT TABLE 3 ABOUT HERE
-------------------------------------------------
In Model 2, we add the diversity of intra-organizational ties. As Hypothesis 1 predicts,
diversity of intra-organizational ties has a positive and significant impact on the likelihood of
achieving high innovative status in this organization. A one standard deviation increase in the
Blau-index capturing diversity of the contacts within Arup increases the innovativeness score of
4
Although the mean and the standard deviation of the dependent variables are very similar, the goodness-
of-fit test comparing the Poisson predictions for a model equivalent to Model 1 in Table 3, indicates that
the Poisson model fits very poorly (χ
2
= 261.92, p=0.000).
5
This assumes that individuals are drawn from a population with a group structure, and that the errors are
correlated within these groups. The clustered error structure compensates for a downward bias that would
result in a model that wrongly assumed no clustered errors.

26
an individual by 19.4%, holding all the other variables constant. This result supports the idea that
drawing resources from other part of the organization enables an individual to benefit by a higher
innovative status with their colleagues. Model 3 includes the diversity of inter-organizational ties.
Contrary to what is predicted in Hypothesis 2, diversity of inter-organizational ties is not a
significant factor in explaining an individual’s innovative status.
In Model 4, we include the variable capturing an individual’s knowledge heterogeneity.
As Hypothesis 3 suggests, individuals with heterogeneous knowledge are more likely to gain
innovative status. An examination of the coefficient estimates suggest that a standard deviation
increase in the knowledge heterogeneity indicator increases the innovativeness score of an
individual by 3.4%, keeping all the other variables constant. In Model 5, we introduce the
interaction effect between knowledge heterogeneity and diversity of intra-organizational ties. The
results of Model 5 show the interaction effect is not positive, but is, in fact, negative and
significant at 1% level. This finding contradicts our Hypothesis 4a and suggests instead that when
the knowledge of an individual is heterogeneous and he/she is drawing resources from colleagues
working in different divisions of the same organization, his/her ability to achieve high innovative
status decreases. In Model 6, we include the interaction effect between knowledge heterogeneity
and diversity of inter-organizational ties. Coefficient estimates for this model do not support
Hypothesis 4b that proposed that knowledge heterogeneity moderates the impact of diversity of
inter-organizational ties on innovative status. Table 4 shows the magnitude of the effects for
Model 5.
-------------------------------------------------
INSERT TABLE 4 ABOUT HERE
-------------------------------------------------
DISCUSSION
Theoretical contributions

27
This paper was motivated by a desire to expand our understanding of why some individuals gain
innovative status within organizations. By theoretically linking a rich body of literature on
external ties to that of knowledge diversity – and empirically substantiate it - our study sought to
contribute to both.
The external face of organizations. Scholars have noted that we need to take greater
appreciation of the character of external ties (Allen, 1977; Cross and Cummings, 2004; Perry-
Smith, 2006). Echoing this view, our paper investigated the character of the ties by focusing on
the diversity of affiliations of these relationships within both the intra- and inter-organizational
network. Using the context of a professional services firm, where professional status is a source
of both advancement and self-realisation for individuals, we found that individuals who possess
diversity of ties with actors from within the organization are likely to gain innovative status.
These individuals are those who their colleagues turn to when they face the most difficult and
challenging aspect of organizational life, the search for new solutions. Such individuals are
valuable for the search because they are able to call upon sources of information that other
individuals with less rich networks are unable to. Accordingly, these individuals gain status from
holding ties with high status individuals in other parts of the organization, tapping in powerful
resources as well as creating strong signals to their colleagues that they are able to connect with
useful people.
Given a rich body of literature on the importance of connecting to external actors (Powell
et al., 1996), we were surprised that diversity of external ties had no effect on an individual’s
innovative status. Some scholars have suggested that individuals inside organizations are
suspicious of external ideas as they may rebel to the existing logic within the social group (Katz
and Allen, 1982). We therefore interpret this as a diversity of external ties may offer dissonant,
challenging and unwanted information and impede an individual’s innovative status.
Going beyond raw counts of external ties to focus on the character of these external ties
helps to unpack the types of external ties that convey internal advantage. It helps to reopen the

28
line of investigation started by Allen and colleagues in the 1960s and 1970s about the interplay
between internal and external networks (Allen, 1984; Tushman, 1977). In doing so, it helps to
provide insight into how individuals and organizations can configure their networks to realize
advantages. Networks with other individuals inside the organization can provide an advantage
when they are diverse, creating opportunities for individuals to access those that others cannot
and to span boundaries which others are unable to. In this sense, the benefits of external
engagement for individuals’ are relational in character, depending on ties of others inside the
organization. Ties to individuals working in external organizations do not appear to affect an
individual’s innovative status. This could be due to the fact that there are often high barriers to
knowledge transfer across organizations. As Szulanski (1996: 27) notes: “because internal
transfers typically are hindered less by confidentiality and legal obstacles than external transfers,
they could be faster and initially less complicated, all other things being equal”.
Knowledge and networks. Implicit in a range of important studies on networks, is an
argument that networks confer knowledge benefits (Burt, 1992). Knowledge heterogeneity needs
to be made theoretically distinct from networks, as opportunities for importing and export ideas
across a network are often missed by those whose experience is limited to single domains (Burt,
2005). This underscores that we need to consider the type of knowledge an individual possesses.
We found that individuals with greater heterogeneity offer something unique in the workplace
and are better positioned to gain innovative status. Extending this view, we contended that
knowledge heterogeneity play a contingent effect on the value of holding diverse intra- and inter-
organizational ties. Contrary to our theoretical argument, we discovered that individuals with less
knowledge heterogeneity benefit more from intra-organizational ties. One potential explanation
for this counter-intuitive finding is that individuals with knowledge heterogeneity already offer
something unique to their colleagues and they may not be able to gain an additional advantage
from interacting with external people. In contrast, individuals that have less knowledge
heterogeneity may become too insular in their knowledge base and can benefit more from having

29
a diversity of ties. An alternative interpretation for this finding is that heterogeneous knowledge
and a diverse intra-organizational network are not complementary, but in fact may be substitutes
for one another.
Taken together, our study offers new insights into a type of brokerage that can accrue
from accessing diverse individuals across organizational boundaries which affects individuals
differently depending on the knowledge they possess.
Managerial implications
The innovative ability of an individual is often highly rewarded within organizations and it is
strongly correlated with positive financial outcomes for individuals. Indeed, extensive managerial
resources are directed to creating conditions to help individuals be more creative and innovative
in their work. Measures that would enable organizations to more successful support
innovativeness among their staff would be highly welcomed by practitioners and could offer an
opportunity to enable organizations to increase the innovativeness of their teams and of their
organization. Initiatives such as P&G’s Connect & Develop have received significant attention
for its recognition to make better use of external scientists outside the four walls of the firm
(Huston and Sakkab, 2006). Our study offers an attempt to conceptualize the diversity of the
affiliations of external partners. In doing so, we aimed to cast light under what circumstances
individuals gain innovative status from external ties. We find little support for the idea that
individuals are automatically better off by searching externally. Indeed, there were no positive
benefits to inter-organizational ties. However, the results do suggest that innovative status goes to
those individuals who can span different departments within the firm and therefore efforts to
promote intra-organizational networking (with other divisions) may be more valuable than
promoting inter-organizational networking.
The importance of knowledge heterogeneity for innovative status also suggests several
managerial implications. The first of these is that understanding the relative character of a
person’s knowledge can provide insights into their potential for innovation. Mapping the

30
knowledge of staff by capturing information in expert pages provides a good opportunity to create
semantic networks of the firm and could offer a means of targeting innovative efforts to those
individuals who span important areas of knowledge. Second, individuals who span different
knowledge domains - the foxes - provide an important source of new ideas to their colleagues,
helping to speed up their search for new solutions. Organizations may seek to actively harness the
skills of these foxes, placing them in cohesive teams to inject vitality to them and using them for
spanning important functions within the organization. Focusing on knowledge heterogeneity may
help to open up a new way of finding and supporting brokers, and therefore in harnessing the
important role of brokerage for supporting innovation in the organization.
Limitations
There are several limitations to this study. Like many other network studies inside organizations,
our data is limited to a single period and therefore it is difficult to draw inferences about the
casual relationship between the main variables. Also we also have information from a single
office of a single firm and therefore our results may reflect the particular patterns of social
relations in Arup. Moreover, our dependent variable – innovative status – comes from the same
source as some of our independent measures. The fact that the survey focuses on two different
network environments – inside and outside the organization – helps dampen fears of common
method bias, but it does not remove it. In addition, our network data on inter-organizational
networks focuses only on five ties and therefore covers only a small share the organization’s
external network. Despite these limitations, which are shared by many studies in this area, the
results suggest a promising line of future research to unpack and explore the interplay between
the internal and external face of the firm’s networks.

31
REFERENCES
Ahuja G. 2000. Collaboration networks, structural holes and innovation: a longitudinal study.
Administrative Science Quarterly 45: 425-455
Aldrich H, Herker D. 1977. Boundary spanning roles and organization structure. Academy of
Management Review 2(2): 217-230
Allen JT. 1984. Managing the Flow of Technology: Technology Transfer and the Dissemination
of Technological Information within the R&D Organization. MIT Press: Cambridge (MA)
Allen T. 1977. Managing the Flow of Technology: Technology Transfer and the Dissemination of
Technological Information Within the R and D Organization. The MIT Press: Cambridge, Mass.
Allen TJ, Cohen SI. 1969. Information flow in research and development laboratories.
Administrative Science Quarterly 14(1): 12-19
Ancona DG, Caldwell DF. 1992. Demography and design: predictors of new product team
performance. Organization Science 3(3): 321-341
Baum JAC, Calabrese T, Silverman BS. 2000. Don't go it alone: Alliance network composition
and startups' performance in Canadian biotechnology. Strategic Management Journal 21: 267-
294
Blau PM. 1977. Inequality and heterogeneity : a primitive theory of social structure. Free Press:
New York
Blyler M, Coff RW. 2003. Dynamic Capabilities, Social Capital, and Rent Appropriation: Ties
That Split Pies. Strategic Management Joumal 24: 677-686
Bonacich P. 1987. Power and Centrality: A Family of Measures. American Journal of Sociology
92(5): 1170-1182
Bonacich P, Lloyd P. 2001. Eigenvector-like measures of centrality for asymmetric relations.
Social Networks 23: 191-201
Borgatti PS, Foster CP. 2003. The Network Paradigm in Organizational Research: A Review and
Typology. Journal of Management 29(6): 991-1114

32
Brass DJ, Burkhardt ME. 1992. Centrality and Power in Organizations. In N Nohria, RG Eccles
(Eds.), Networks and Organizations: 191-215. Harvard Business School Press: Boston, MA
Brass DJ, Galaskiewicz J, Greve HR, Tsai W. 2004. Taking stock of networks and organizations:
a multilevel perspective. Academy of Management Journal 47(6): 795-817
Burt RS. 1992. Structural holes: The social structure of competition: Cambridge, MA.
Burt RS. 2004. Structural holes and good ideas. American Journal of Sociology 110(2): 349-399
Burt RS. 2005. Brokerage & Closure: An Introduction to Social Capital. Oxford University
Press: Oxford
Chesbrough H. 2003. Open innovation. Harvard University Press: Cambridge, Massachusetts
Cohen WM, Levinthal DA. 1990. Absorptive capacity: a new perspective on learning and
innovation. Administrative Science Quarterly 35(1): 128-152
Criscuolo P, Salter A, Sheehan T. 2007. Making Knowledge Visible: Using Expert Yellow Pages
to Map Capabilities in Professional Services Firms. Research Policy Forthcoming
Cross R, Cummings JN. 2004. Tie and network correlates of individual performance in
knowledge-intensive work. Academy of Management Journal 47(6): 928-937
Davies A, Hobday M. 2005. The Business of Projects: Managing Innovation in Complex
Products and Systems. Cambridge University Press: Cambridge
Dearborn DC, Simon HA. 1958. Selective Perception: A Note on the Departmental Identification
of Executives. Sociometry 21: 140-144
Dodgson M, Gann DM, Salter A. 2007. 'In case of fire, please use the elevator': simulation
technology and organization in fire engineering Organization Science 18(5): 849-864
Dougherty D. 2004. Organizing practices in services: capturing practice-based knowledge for
innovation. Strategic Organization 2(1): 35-64
Evans JA, Barley SR, Kunda G. 2004. Beach time, bridge time, and billable hours: The temporal
structure of technical contracting. Administrative Science Quarterly 49: 1-38

33
Fleming L, Sorenson O. 2004. Science as a map in technological search. Strategic Management
Journal 25(8-9): 909-928
Freeman LC. 1979. Centrality in social networks. Social Networks 1: 215-240
Geletkanycz MA, Hambrick DC. 1997. The External Ties of Top Executives: Implications for
Strategic Choice and Performance. Administrative Science Quarterly 42(4): 654-681
Hansen M. 2000. Knowledge networks: Explaining effective knowledge sharing in multiunit
companies. Organization Science 13: 232-248
Hansen MT. 1999. The Search-Transfer Problem: The Role of Weak Ties in Sharing Knowledge
across Organization Subunits. Administrative Science Quarterly 44(1): 82-111
Hargadon A, Sutton R. 1997. Technology brokering and innovation in a product development
firm. Administrative Science Quarterly 42(4): 716-749
Hayek FAv. 1948. The Use of Knowledge in Society, Individualism and Economic Order: 77-91.
University of Chicago Press: Chicago
Hitt M, Bierman L, Shimizu K, Kochhar R. 2001. Direct and Moderating Effects of Human
Capital on Strategy and Performance in Professional Service Firms: A Resource-Based
Perspective. Academy of Management Journal 44(1): 13-28
Huston L, Sakkab N. 2006. Connect & Develop: Inside Procter & Gamble's New Model of
Innovation. Harvard Business Review March: 58-68
Ibarra H. 1993. Network Centrality, Power and Innovational Involvement: Determinants of
Technical and Administrative Roles. Academy of Management Journal 36(3): 471-501
Katz R, Allen T. 1982. Investigating the Not Invented Here (NIH) syndrome: A look at the
performance, tenure, and communication patterns of 50 R&D projects. R&D Management 12(1):
7-19
Kogut B, Zander U. 1992. Knowledge of the firm, combinative capabilities, and the replication of
technology. Organization Science 3(3): 383-397

34
Laursen K, Salter AJ. 2006. Open for Innovation: The role of openness in explaining innovative
performance among UK manufacturing firms. Strategic Management Journal 27(2): 131-150
Lazega E. 2000. Rule Enforcement among Peers: A Lateral Control Regime. Organization
Studies 21: 193-214
Lazega E. 2001. The Collegial Phenomenon. Oxford University Press: Oxford
Maister D. 1993. Managing Professional Service Firms. Free Press: New York
Majchrzak A, Cooper LP, Neece OE. 2004. Knowledge Reuse for Innovation. Management
Science 50(2): 174-188
March J. 1991. Exploitation and Exploration in Organizational Learning. Organization Science
2(1): 71-87
March JG, Simon HA. 1958. Organizations: NY. US.
Marcson S. 1960. The Scientist in American Industry. . Princeton University: Princeton: Industrial
Relations Section
McKenna P, Maister D. 2002. First Amongst Equals. Free Press: New York
Menon T, Pfeffer J. 2003. Valuing internal vs. external knowledge: Explaining the preference for
outsiders. Management Science 49(4): 497-513
Monteiro F, Arvidsson N, Birkinshaw J. 2007. Knowledge flows in multinational corporations.
Organization Science forthcoming
Moulton BR. 1990. An illustration of a pitfall in estimating the effects of aggregate variables on
micro units. The Review of Economic Statistics 72(2): 334-338
Obstfeld D. 2005. Social Networks, the Tertius Iungens Orientation, and Involvement in
Innovation. Administrative Science Quarterly 50(1): 100-130
Padgett JF, Ansell CK. 1993. Robust action and the rise of the Medici, 1400-1434. American
Journal of Sociology 98(6): 1259-1319

35
Perry-Smith JE. 2006. Social yet creative: The role of social relationships in facilitating
individual creativity. Academy of Management Journal 49(1): 85-101
Podolny JM, Baron JN. 1997. Resources and relationships: Social networks and mobility in the
workplace. American Sociological Review 62(5): 673-693
Podsakoff PM, MacKenzie SB, Lee J-Y, Podsakoff NP. 2003. Common method bias in
behavioral research: A critical review of the literature and recommended remedies. Journal of
Applied Psychology 88(5): 879-903
Powell WW, Koput KW, Smith-Doerr L. 1996. Interorganizational Collaboration and the Locus
of Innovation: Networks of Learning in Biotechnology. Admistrative Science Quarterly 41(1):
116-145
Reagans R, McEvily B. 2003. Network Structure and Knowledge Transfer: The Effects of
Cohesion and Range. Administrative Science Quarterly 48: 240-267
Rodan S, Galunic C. 2004. More Than Network Structure: How Knowledge Heterogeneity
Influences Managerial Performance and Innovativeness. Strategic Management Joumal 25: 541-
562
Rollag K. 2004. The impact of relative tenure on newcomer socialization dynamics. Journal of
Organizational Behavior 25: 853-872
Saxenian A. 1994. Regional Advantage: Culture and Competition in Silicon Valley and Route
128. Harvard University Press: Cambridge, MA
Schumpeter JA. 1912/34. The Theory of Economic Development: An Inquiry into Profits, Capital,
Credit, Interest and the Business Cycle (R Opie, Trans.). Oxford University Press: London
Stuart T, Podolny J. 1996. Local search and the evolution of technological capabilities. Strategic
Management Joural 17(Special Issue: evolutionary perspectives on strategy (Summer)): 21-38
Szulanski G. 1996. Exploring internal stickiness: impediments to the transfer of best practice
within the firm. Strategic Management Journal 17(Winter Special Issue): 27-43
Teece DJ. 2003. Expert talent and the design of (professional services) firms. Industrial and
Corporate Change 12(4): 895-916

36
Thompson JD. 1967. Organizations in action. McGraw-Hill: New York, US.
Tortoriello M, McEvily B, Krackhardt D. 2008. Disentangling social structure and knowledge
structure: external knowledge, social networks, and individual innovativeness, HEC working
paper: Paris
Tsai W. 2001. Knowledge transfer in intraorganizational networks: effects of network position
and absorptive capacity on business unit innovation and performance. Academy of Management
Journal 44(5): 996-1004
Tsai WP. 2002. Social structure of "coopetition" within a multiunit organization: Coordination,
competition, and intraorganizational knowledge sharing. Organization Science 13(2): 179-190
Tushman ML. 1977. Special boundary roles in the innovation process. Administrative Science
Quarterly 22: 587-605
Tushman ML, Scanlan TJ. 1981. Boundary spanning individuals: Their role in information
transfer and their antecedents. Academy of Management Journal 24: 289-305
Van Maanen J, Barley SR. 1984. Occupational communities: Culture and control in
organizations. Research in Organizational Behavior 6: 287-365
Veshosky D. 1998. Managing innovation information in engineering and construction firms.
Journal of Management in Engineering 14(1): 58-66
Wasserman S, Faust K. 1994. Social Network Analysis: Methods and Applications Cambridge
University Press London

37
Table 1. Survey items for the social network questions
Network type
Name generator questions
Problem Solving
Some people are particularly useful in
helping you solve problems in your work
activities (e.g. a simulation software tool does
not work, you cannot find a solution to a
formula that applies to a particular model,
you cannot resolve a contractual issue). Who
are the key people in the office who have
helped you solve problems in the last 6
months by defining or redefining the
dimensions of the problem and identifying
relevant information?
Search
Some contacts are particularly useful in
helping you to be creative in your job, such
as helping you to explore new engineering
solutions, ideas or concepts. Who are the key
people in the office that have helped you
formulate new ideas during the past 6
months?
Implementation
Often support from others may be needed to
validate and implement your proposed
solutions. Who are the key people in the
office that provided support to you in
validating and implementing your solutions
during the past 6 months?
Advice
Think of people who acted as a critical source
of knowledge for your projects during the
past 6 months These are people who you
collaborate with or turn to for advice when
you needed help solving a problem
encountered in one of your projects.

38
Table 2 Descriptive statistics and correlation matrix of variables
Variable Mean Std. Dev. Min Max 4 7 9 11 13
1 Innovative Status 5.21 5.39 0 28
2 Junior 0.21 0.41 0 1 -0.31
*
3 Experienced 0.24 0.43 0 1 -0.25
*
-0.28
*
4 Senior 0.35 0.48 0 1 -0.07 -0.37
*
-0.41
*
5 Director 0.21 0.41 0 1 0.64
*
-0.25
*
-0.28
*
-0.37
*
6 Tenure 7.06 6.20 1 28 0.36
*
-0.37
*
-0.11 -0.07 0.56
*
7 Full time 0.94 0.24 0 1 0.10 0.05 0.07 -0.15 0.05 0.07
8 Master 0.36 0.48 0 1 0.24
*
-0.18
*
-0.08 0.05 0.21
*
0.24
*
0.06
9 PhD 0.05 0.21 0 1 0.09 0.07 -0.04 0.07 -0.11 -0.08 0.06 -0.16
10 Member Professional Organization 0.34 0.48 0 1 0.21
*
-0.37
*
-0.14 0.11 0.39
*
0.35
*
0.05 0.13 -0.08
11 Gender 0.76 0.43 0 1 0.13 -0.02 -0.10 0.01 0.11 0.06 0.15 0.17 0.12 0.22
*
12 Skill Description Length 454.96 289.56 3 1193 0.16 -0.17 0.05 -0.01 0.13 0.17 0.07 0.13 0.09 0.04 0.05
13 Knowledge Heterogeneity 0.51 0.65 0 3.38 0.29
*
-0.19
*
0.02 -0.06 0.24
*
0.22
*
0.10 0.16 0.12 0.13 0.14 0.52
*
14 Diversity Intra-organizational Ties 0.36 0.32 0 0.89 0.32
*
-0.20
*
-0.18
*
0.09 0.28
*
0.21
*
0.11 0.16 -0.04 0.18
*
-0.04 0.05 0.16
15 Diversity Inter-organizational Ties 0.30 0.32 0 0.8 0.29
*
-0.22
*
-0.06 0.03 0.26
*
0.13 0.06 0.18
*
-0.07 0.15 -0.05 0.15 0.17 0.19
*
8 10 12 141 2 3 65
*
p<0.05

Table 3 Negative binomial model estimation results
1 2 3 4 5 6
Constant 1.989 2.515 2.523 2.54 2.423 1.869
(0.462)
***
(0.294)
***
(0.289)
***
(0.296)
***
(0.351)
***
(0.361)
***
Junior -1.995 -1.908 -1.913 -1.909 -1.852 -1.916
(0.069)
***
(0.054)
***
(0.076)
***
(0.084)
***
(0.143)
***
(0.090)
***
Experienced -1.493 -1.351 -1.354 -1.357 -1.325 -1.364
(0.059)
***
(0.067)
***
(0.082)
***
(0.087)
***
(0.124)
***
(0.092)
***
Senior -1.043 -0.973 -0.975 -0.967 -0.925 -0.966
(0.094)
***
(0.072)
***
(0.068)
***
(0.073)
***
(0.125)
***
(0.070)
***
Tenure -0.006 -0.005 -0.005 -0.005 -0.002 -0.005
(0.007) (0.005) (0.005) (0.005) (0.004) (0.005)
Full-time 0.79 0.698 0.702 0.705 0.637 0.716
(0.261)
***
(0.238)
***
(0.212)
***
(0.207)
***
(0.216)
***
(0.191)
***
Master 0.311 0.297 0.298 0.293 0.247 0.293
(0.226) (0.176)
*
(0.172)
*
(0.173)
*
(0.189) (0.172)
*
PhD 0.638 0.652 0.649 0.622 0.628 0.617
(0.306)
**
(0.242)
***
(0.240)
***
(0.238)
***
(0.237)
***
(0.230)
***
Member Professional
Organization -0.119 -0.151 -0.151 -0.156 -0.141 -0.161
(0.141) (0.141) (0.142) (0.143) (0.161) (0.151)
Gender 0.002 0.007 0.007 0 0.009 -0.009
(0.174) (0.139) (0.141) (0.136) (0.126) (0.140)
Skill Description
Length 0.001 -0.005 -0.005 -0.01 -0.006 -0.01
(0.007) (0.006) (0.006) (0.006)
*
(0.007) (0.006)
*
H1: Diversity Intra-
organizational Ties 0.555 0.557 0.546 0.597 0.542
(0.143)
***
(0.132)
***
(0.135)
***
(0.163)
***
(0.134)
***
H2: Diversity Inter-
organizational Ties -0.019 -0.023 0.002 -0.038
(0.176) (0.189) (0.175) (0.211)
H3: Knowledge
Heterogeneity 0.056 0.114 0.047
(0.027)
**
(0.067)
*
(0.017)
***
H4a: Knowledge
Heterogeneity X
Diversity Intra-
organizational Ties -0.382
(0.095)
***
H4b: Knowledge
Heterogeneity X
Diversity Inter-
organizational Ties 0.083
(0.114)
Observations 131 131 131 131 131 131
Log-likelihood -307.03 -303.98 -303.98 -303.86 -302.72 -303.81
Alpha 0.22 0.19 0.19 0.19 0.18 0.19
Robust clustered errors on grade classes in brackets. * significant at 10%; ** significant at 5%;
*** significant at 1%.

Table 4. Magnitude of estimated effects from negative binomial regressions
Variables
% Change in expected count for
Std.Dev. increase in X
Junior -52.9
Experienced -43.2
Senior -35.8
Tenure -1
Full time 16.5
Master 12.6
PhD 14.1
Member Professional Organizations -6.5
Gender 0.4
Skill Description Length -1.7
Diversity Intra-organizational Ties 21
Diversity Inter-organizational Ties 0.1
Knowledge heterogeneity 7.7
Knowledge heterogeneity x Diversity Intra-
organizational Ties -8.2
Note: This table reports the percentage change in the dependent variable associated with a one
standard deviation change in a given independent variable, keeping all the variables included
in model 5 constant. For the dummy variables, the percentage change in the dependent
variable associated when the dummy variable changes from 0 to 1, keeping the other
independent variables constant