Network Brokerage:
How the Social Network Around You
Creates Competitive Advantage
for Innovation and Top-Line Growth
For text on this session,
see Chapters 1 and 2 in
Brokerage and Closure
(including adjunct bits
from Neighbor Networks).
Appendices:
Strategic Leadership
Network Brokerage (page 1)
I. Example Network Questionnaire for a Web Survey (pages 29-30, from 2010, Neighbor Networks)
II. Network Metrics (pages 31-37, from 1992, Structural Holes, and 2010, Neighbor Networks)
III. NetDraw Quick Start -- making your own sociograms and benchmark network metrics (page 38)
IV. Network Endogeneity -- Bavelas-Smith-Leavitt experiments (pages 39-40, 1949 Leavitt dissertation, 1951 Leavitt,
“Some effects of certain communication patterns upon group performance”)
V. National Differences in Business Culture (pages 41-44)
This handout was prepared as a basis for discussion in executive education (Copyright © 2016 Ronald S. Burt, all rights reserved). To download
work referenced here, or research/teaching materials on related topics, go to http://faculty.chicagobooth.edu/ronald.burt.
Sociogram of the Org Chart for a
Large EU Healthcare Organization
CEO
C-Suite
Heir Apparent
Other, Respondent
Other, NonRespondent
Bill
Strategic Leadership
Network Brokerage (page 2)
Bob
Figure 1.1 in Burt (2017, Structural Holes in Virtual Worlds).
Sociogram of Senior Leadership in the
Healthcare Organization
Asia
US
Lines indicate frequent and
substantive work discussion;
heavy lines especially
close relationships.
Bill
EU and
Emerging
Markets
Bob
B
Front
Office
B
B
B
Strategic Leadership
Network Brokerage (page 3)
B
B
CEO
C-Suite
Heir Apparent
Back
Office
B
R&D
Other Senior Person
Figure 1.2 in Burt (2017, Structural Holes in Virtual Worlds).
Coordination across groups is the source of competitive advantage from social
networks — an advantage often difficult to realize, as illustrated by the below
opinions from a senior executive education program. The difficulty most often
cited as needing to be improved: Coordination across groups.
SURVEY QUESTION: Every person in the [company] program has a perspective on [company] operations and clients. Drawing on
your experience, what would you like to see changed within [the company] to increase the company's current value as a provider
of high-quality, effective, attractive service to clients? RESPONSES: At least two-thirds of responses are variations on "improve
coordination across groups." For example (idea ratings are based on independent pile-sort evaluations by the head of HR and a regional P&L leader):
HIGH RATED: We need a "DNA" change - we are so deeply ingrained in the P&L, near-term financial results mindset that we limit
our potential for changing the world. It goes beyond P&L reporting or annual goals - we need more people talking and thinking and
acting and making decisions that are focused on one voice, delivering for clients and colleagues. We are not a "clients first" firm we say we are, but we are not. We are a Wall Street/investor first firm. Need to change that.
Strategic Leadership
Network Brokerage (page 4)
HIGH RATED: (1) Improve team work in global network. Global clients want to see us acting more as one team (e.g. global P&L for
TOP Global clients). (2) Better understanding of capabilities and solutions of other business units, and improving colleague network
cross business units (one voice). (3) Best practice sharing cross countries, esp. for same roles/job titles to copy/paste success
stories and solutions for our clients. (4) Improve incentive/bonus structure to drive x-Selling (in country between BUs, also globally
within BUs). (5) Better understanding of [company]'s global capabilities and whom to talk to for a special solution.
HIGH RATED: (1) “One Client View." Ability to easily understand [company]-wide relationships with any client or prospect in the
world. Comprehensive, single instance client account data system. Ability to pull up information on a client (any business unit) from
one system. Data should include: revenue, oppty amount, current solutions provided, account owner(s), history of account, etc. (2)
Connect the dots between business initiatives and commitments, and finance/budget. When the business makes commitments to
do something, sometimes we need to do a better job at ensuring those commitments are resourced appropriately (otherwise we end
up with frustrated and over-worked colleagues with compensation not commensurate with their scope and/or efforts).
ILLUSTRATIVE RELATED IDEAS:
- More coordination among different parts of the company. A better understanding of each other's business and the value it
brings to clients. A clearer internal structure that allows us to deliver to clients the best of [company] with internal P&L driven issues.
- Continue to endeavor to remove our P&Ls as barriers to cross-practice collaboration.
- A far more joined-up [company]. We don't speak to client yet routinely about our true capability across [company]. We are
moving in that direction, but this would be a genuine game changer and add material value. Not a new idea I know!
- Better integrated client management activities across the operating groups.
- Find an easier way to work across businesses and geographies to really bring the best of [company] to all clients.
To begin, the "network" around a person is a pattern of
relationships with and between colleagues.
This worksheet is completed in four steps:
(1) In the oval, write the name of a significant
colleague. The colleague could be your most
valuable subordinate, your most difficult, your
boss, an important source of support, or a
key contact in another organization. Who
doesn't matter. It just has to be someone you
know well enough to know their key contacts.
(2) In the squares, write the
name of the five contacts with
whom the person in the oval
has the most frequent and
substantial business contact.
Strategic Leadership
Network Brokerage (page 5)
(3) Draw a line between any pair of contacts
that are connected in the sense that the two
people speak often enough that they have
some familiarity with current issues in one
another's work.
(4) Compute network density. Count the
number of lines between contacts (TIES).
Divide by the number possible (n[n-1]/2,
where n is the number of contacts, which is 5
if you entered five contacts). Multiply by 100
and round to nearest percent.
DENSITY = _____________
Appendix I contains an illustrative survey webpage used to gather network data.
SOCIOGRAM
graphic image of
a network in which
dots represent
nodes (a person,
group, etc.) and
lines represent
connections
Social Capital of Brokerage
2.5
Manifest as better ideas, more-positive evaluations, higher compensation,
earlier promotion, and faster teams.
Z-Score Relative Performance
(compensation, evaluation, promotion)
2.0
Now, establish the
empirical fact that
the people we will
discuss as "network
brokers" enjoy
achievement and
rewards higher than
their peers. Brokers
are distinguished on
the horizontal axis.
1.5
1.0
0.5
0.0
-0.5
-1.0
-1.5
Strategic Leadership
Network Brokerage (page 6)
-2.0
-2.5
5
15
25
35
45
55
65
75
85
95
Network Constraint (C)
large, open
many ——— Structural Holes ——— few
Circles are average z-score performance (Z) for a five-point interval of network constraint
Robert
(C) within each of eight study populations. Dashed line goes through mean values of Z for
intervals of C. Bold line is performance predicted by the natural log of C.
small, closed
James
From Figure 1.8 in Brokerage and Closure. Data pooled across eight
study-population graphs in Appendix II on measuring network constraint.
E
(compensation, evaluation, promotion)
Z-Score Relative Performance
E
1.5
E
1.0
J
E
0.5
E
median network
constraint (35 points)
E
E
E
performance
E
E EE
lower than
E E E
EE
E
E
E
E
expected
E
E
E E
E
E
E
E E
J
J
E
J
J E J
JE E
J
E
J
E
E
E E
E
J
E
J
E
J
J
E
E
Human Capital et Eal.
E
EEgeography, kind of work, division, education, etc.)
EE
(e.g., job
rank, age,
E
E
E
E
E
E
E
E E
0.0
-0.5
-1.0
-1.5
-2.5
E
performance
than more-positive evaluations, higher compensation,
Manifest as higher
better ideas,
expected
earlier promotion, and faster teams.
E
E J
E
EE E
JE E E EE
E
J E
E
E E
E
E
E
E
E
J
E
EE
E E
E
E E
E J
E EE E
-2.0
Strategic Leadership
Network Brokerage (page 7)
Performance Indicator
E
2.0
(compensation, evaluation, promotion rate)
2.5
Social Capital of Brokerage
E
E
E
E
5
15
25
35
E
45
55
65
75
Z = 2.78 - .82 ln(C)
r = -.53
85
95
Achievement
and rewards are
distinguished
on the vertical
axis,
measuring the
extent to which
a person is
doing better
than his or her
peers.
Network Constraint (C)
large, open
many ——— Structural Holes ——— few
Circles are average z-score performance (Z) for a five-point interval of network constraint
Robert
(C) within each of eight study populations. Dashed line goes through mean values of Z for
intervals of C. Bold line is performance predicted by the natural log of C.
small, closed
James
From Figure 1.8 in Brokerage and Closure. Data pooled across eight
study-population graphs in Appendix II on measuring network constraint.
Social Capital of Brokerage
2.5
Manifest as better ideas, more-positive evaluations, higher compensation,
earlier promotion, and faster teams.
E
(compensation, evaluation, promotion)
Z-Score Relative Performance
Brokerage is a large percentage of explained performance differences.
E
2.0
E
1.5
E
1.0
E
E
J
E
E
E J
E
EE E
JE E
E E EE
E
J
E E
E E
E
E
E
E
E
J
E
E
EE
E E
E
E E
E
EJ
E EE
0.5
0.0
-0.5
E
E
E EE
E
E E
E
E
E
E
E
J E J J
E
E
E
EE
EE
-1.5
E
E
E E
E
E
E
E
J
J E
E
17%
EE
J
E
E
J
E
E
J
J E
E
E
EE
J
J
E
E
9%
E
E
E
E
E
E
5
15
25
35
E
45
55
E
E
81%
65
75
Z = 2.78 - .82 ln(C)
r = -.53
85
95
Network Constraint (C)
large, open
10%
E
J
E E
-2.0
Strategic Leadership
Network Brokerage (page 8)
E
E
-1.0
-2.5
E
55%
28%
median network
constraint (35 points)
many ——— Structural Holes ——— few
Circles are average z-score performance (Z) for a five-point interval of network constraint
Robert
(C) within each of eight study populations. Dashed line goes through mean values of Z for
intervals of C. Bold line is performance predicted by the natural log of C.
33%
2%
small, closed
James
64%
Brokerage
Contributes
"Slightly More
than Half"
of Predicted
Variance in
Performance
Differences
between
Managers:
Network
constraint (white),
job rank (red),
and other factors
(striped). First
pie is investment
banker
compensation
and analyst
election to the
All-America
Research Team.
Second pie is
supply-chain and
HR manager
compensation
in corporate
bureaucracies.
Third pie is
early promotion
to senior job
rank in a large
electronics firm.
Graph is from Figure 1.8 in Brokerage and Closure. Data are pooled
across eight management populations. Pie charts are from Figure
2.4 in Neighbor Networks. On causal order, see Appendix IV.
Returns to Brokerage Aggregate
to Companies, Industries, and Communities
People with phone networks
that span structural holes
live in communities higher
in socio-economic rank
Strategic Leadership
Network Brokerage (page 9)
Networks are defined by land-line & mobile
phone calls (map to left). Socio-economic
rank is UK government index of multiple
deprivation (IMD) based on local income,
employment, education, health, crime,
housing, and environmental quality (graph
below). Units are phone area codes.
figures from Eagle, Macy, and Claxton (2010, Science), “Network diversity and economic development”
Returns to Brokerage Are Evident
in Low Returns to Pre-Job Specialization
Recent scholarship on the returns to labor market specialization often claims
that being specialized is advantageous for job candidates. We argue, in contrast,
that a specialist discount may occur in contexts that share three features: strong
institutionalized mechanisms, candidate profiles with direct investments that
signal their value, and a high supply of focused candidates relative to demand.
We then test whether there is a specialist discount for graduating elite MBAs,
as it is a labor market that exemplifies these conditions under which we expect
specialists to be penalized. Using rich data on two graduating cohorts from a toptier U.S. business school (full-time students, 2008-2009), we show that elite MBA
graduates who established a focused (specialized) market profile of experiences
relating to investment banking before and during the program were less likely to
receive multiple job offers and were offered less in starting-bonus compensation
than similar MBA candidates with no exposure or less-focused exposure to
investment banking. Our theory and findings suggest that the oft-documented
specialist advantage may be overstated.
Strategic Leadership
Network Brokerage (page 10)
Figure 1 displays predicted (marginal) probabilities of receiving multiple offers for
candidates who have mean values for each of the control variables but different
profiles.
Figure 2 compares the starting bonuses of hypothetical job candidates with different
profiles. Each hypothetical candidate is a single white male who graduated from a
top-20 undergraduate institution, has above a 3.8 GPA, received more than one
job offer, has the mean age and work experience characteristics (months, number
of firms), accepts a job in I-banking, and earns the mean base salary for I-banking
jobs in his 2008 cohort year. The only difference is the candidate’s profile in terms
of exposure to I-banking.
FOCUSED (career history in finance before mba, concentration in finance, joined
an i-banking club during mba, and i-banking internship; 61% of students who
graduate to a job in i-banking were focused on i-banking)
NON-SEQUENTIAL exposure (neither of the above categories, but some mba
program contact with i-banking)
PARTIAL sequential exposure (prior experience in finance + concentration in
finance or participation in i-banking club)
PRE-MBA exposure (only exposure before mba program)
figures and text from Merluzzi and Phillips (2016, Administrative Science Quarterly), “The Specialist Discount,"
and for more applied discussion, see Merluzzi, (June 2016, HBR), "Generalists get better job offers than specialists."
Strategic Leadership
Network Brokerage (page 11)
Second Life prediction from
constrained social network
Second Life prediction from
nonredundant social contacts
EverQuest II prediction from
nonredundant social (upper)
versus economic (lower) contacts
EverQuest II prediction
from constrained social (upper)
versus economic (lower) network
25+
Effective Size
(Number of NonRedundant Contacts)
Network Constraint (x 100)
This is Figure 3.5 in Burt, Structural Holes in Virtual Worlds (2017). Dots are average Y scores within integer (left) or five-point (right) intervals on
horizontal axis. EverQuest II achievement variable is the predicted character level in Model 8, Tables 3.4 and 3.5. Second Life achievement is the
canonical correlation dependent variable in Model 15, Tables 3.5 and 3.6 (associations with individual achievement dimensions in Second Life are given
in graphs in Appendix V).
Predicted Avatar Z-score Achievement
Predicted Avatar Z-score Achievement
Returns to Brokerage Are Evident Online in the
Network-Achievement Connection within Virtual Worlds
And the Network Brokerage Effect Is Evident in
Coordination between Organizations
Strategic Leadership
Network Brokerage (page 12)
Supplier Evaluation of Telecom
(mean eval of forecast accuracy and development cycle volatilty)
(1 = unacceptable, 2 = satisfactory, 3 = meets requirement
Supplier evaluation: Leaders in 32 key supplier
organizations were interviewed about their
experience with a leading American telecom
("the company"). For larger suppliers, two or
three agents were interviewed (e.g., Foxconn,
Samsung, Sharp, Toshiba). The agents were
asked to describe their experience with respect
to company forecast accuracy and volatility in
the company's development cycle (3 meets
requirements, 2 satisfactory, or 1 unacceptable).
The vertical axis is the average evaluation of the
telecom company from each key supplier.
r = -.44
t = -2.89
P < .01
Network Constraint on Best-Connected
Procurement Manager Assigned to the Supplier
(lowest network constraint score among managers for whom
the supplier is where manager spends the most time)
Supplier POC in telecom: The 55 managers in
procurement support (no direct supplier contact)
were asked to indicate their involvement in
company operations with each of the 32 key
suppliers. A manager could say that a supplier
is one "on which I spend the most time," or
"with which I have some direct contact," or "on
which I work indirectly through other motorola
employees" (or leave it blank if respondent had
no contact with the supplier). For each of the 32
key suppliers, I identified the respondents who
said they spend "most time" on the supplier, and
selected the two respondents who had the most
attractive network metrics defined by "frequent
and substantive work contact" relations. The
horizontal axis is the average network constraint
score for the two best-connected procurement
managers spending "most time" on the supplier
providing the evaluation on the vertical axis.
HOW IT WORKS: Recombinant Sticky Information
Contacts as Source vs. Portal
YOU
YOU
YOU
Network A
Network B
Network C
Strategic Leadership
Network Brokerage (page 13)
Redundancy
by Cohesion
YOU
A
Contact
Redundancy
1
B
7
2
3
James
Robert
Redundancy
by Structural
Equivalence
5
YOU
4
6
C&D
Network Constraint
(C = Σj cij = Σj [pij + Σq piqpqj]2, i,j ≠ q)
85
Density Table
5
25
0
1
100
0
0
29
Group A
Group B
Group C
0
Group D
person 3: .402 = [.25+0]2 + [.25+.084]2 + [.25+.091]2 + [.25+.084]2
Robert: .148 = [.077+0]2 + [.154+0]2 + [.154+0]2 + [.154+0]2 + [.154+0]2 + [.154+0]2 + [.154+0]2
from Figures 1.1 and 1.3 in Burt (1992, Structural Holes) and Figure 1.2 in Brokerage and Closure
Bridge & Cluster: Small World of Organizations & Markets
A
1
B
7
2
3
James
Robert
5
4
Strategic Leadership
Network Brokerage (page 14)
6
C&D
Network Constraint
(C = Σj cij = Σj [pij + Σq piqpqj]2, i,j ≠ q)
85
Density Table
5
25
0
1
100
0
0
29
Group A
Group B
Group C
0
Group D
person 3: .402 = [.25+0]2 + [.25+.084]2 + [.25+.091]2 + [.25+.084]2
Robert: .148 = [.077+0]2 + [.154+0]2 + [.154+0]2 + [.154+0]2 + [.154+0]2 + [.154+0]2 + [.154+0]2
Network
indicates
distribution
of sticky
information,
which defines
advantage.
From Figure 1.1 in Brokerage and Closure. For an HBR treatment of the network distinction between Robert and James, see in the course packet Kotter's classic distinction
between "leaders" versus "managers." Robert ideally corresponds to the image of a "T-shaped manager," nicely articulated in Hansen's HBR paper in the course packet.
Here is the core network for a job BEFORE and AFTER the employee
expanded the social capital of the job by reallocating network time and
energy to more diverse contacts.
Create Value
by Bridging
Structural
Holes
It is the weak contact connections (structural holes) in
the AFTER network that provides the expanded social
capital.
1
2
STRUCTURAL HOLE
disconnection between two
groups or clusters of people
Strategic Leadership
Network Brokerage (page 15)
BRIDGE
relation across structural hole
NETWORK ENTREPRENEUR
or "broker," or "connector:"
a person who coordinates
across a structural hole
BROKERAGE
act of coordinating across
a structural hole
4
5
.6 con
s
t
rai
nt
STICKY INFORMATION
Information expensive to move
because: (a) tacit, (b) complex,
(c) requires other knowledge to
absorb, or (d) interaction with
sender, recipient, or channel.
53
BEFORE
3
The employee AFTER is more positioned
at the crossroads of communication
between social clusters within the firm
and its market, and so is better
positioned to craft projects and
policy that add value across
clusters.
Research shows that
employees in networks
like the AFTER network,
spanning structural holes,
are the key to integrating
20
operations across functional
.0
and business boundaries. In
co
ns
tra
research comparing senior people
int
*
with networks like these BEFORE and
AFTER networks, it is the AFTER networks
that are associated with more creativity, faster
learning, more positive individual and team
evaluations, faster promotions,
and higher earnings.
information
breadth,
timing, and
arbitrage
1
2
AFTER
*Network scores refer to direct contacts.
From Figure 1.4 in Burt (1992, Structural Holes)
and Figure 1.2 in Brokerage and Closure. See Appendix I on survey network data, Appendix II on measuring network constraint.
3
4
5
Illustrative Networks
around Early and
Late Adopters of the
Chinese Social Media
Service Weibo
Strategic Leadership
Network Brokerage (page 16)
Sociograms show connections among the
people and organizations a user follows.
The network around the early adopter
reveals a broker across three social
clusters. In comparison, the late adopter
is embedded in a single closed network.
Released in 2009, following suppression
of western microblogging services
such as Twitter, Facebook, Plurk and
Fanfou, Weibo is a social media service
that combines elements of Twitter and
Facebook. It is one of the largest such
services in the world (over half a billion
users at the end of 2012).
Figures are from Burt, Huang, Tang, and
Zhang, "Sampling Weibo" (2016).
Guangdong
Late User
This is a male 9.90 months
in Weibo posting 1,916
messages and following
81 others during the observation
period (73.0 nonredundant contacts,
10% mutual contacts, 1 early adopter,
79% in Guangdong, 67% celebrities, 13.6% network density among
contacts (50% mutual), 4.5 constraint from contacts, 1,774.1 betweenness).
Symbols are people the user followed
during the observation period (10
additions during the period). Circles
are contacts who also live in
Guangdong. Squares are contacts who
live elsewhere. Larger symbols are
contacts with whom the user has a
mutual relationship (each follows the
other). Lines indicate connections
between the user’s contacts.
Heavy lines indicate mutual
relations. Thin lines
indicate ties
in which only
one follows
the other.
Symbols are people the user followed during the
observation period (7 changed during the period). Circles
are contacts who also live in Guangdong. Squares are
contacts who live elsewhere. Larger symbols are contacts
with whom the user has a mutual relationship (each
follows the other). Lines indicate connections
between the user’s contacts. Heavy lines
indicate mutual relations. Thin lines
indicate connections in which
only one contact follows
the other.
Common user (10)
Star user (15)
Celebrity user (54)
Organization user (2)
Common user (10)
Star user (15)
Celebrity user (54)
Organization user (2)
Guangdong
Early User
This is a male
32.93 months
in Weibo, posting
3,805 messages,
and following 198 contacts during the
observation period (187.5 nonredundant
contacts, 13% mutual, 11 early adopters,
17% Guangdong, 20% celebrities, 8.3% density
(22% mutual), 2.3 constraint, 9,742.4 betweenness).
HOW IT WORKS: Creativity and Innovation
Are at the Heart of It
Achievement & Rewards
(What benefits?)
Brokerage
across
Structural Holes
What in your work
improves the odds
(How to frame it & who should be involved?)
that you will discover
the value of something
you don't know you don't know?
Adaptive Implementation
Creativity & Innovation
Strategic Leadership
Network Brokerage (page 17)
(What should be done?)
Alternative Perspective (how would this problem look from the perspective of
a different group, or groups — thinking “out of the box” is often less valuable than seeing
the problem as it would look if you were inside a specific “other box”)
Best Practice (something they think or do could be valuable in my operations)
Analogy (something about the way they think or behave has implications for how I can enhance the value of my operations; i.e., look for the value of
juxtapositioning two clusters, not reasons why the two are different so as to be irrelevant to one another — you often find what you look for)
Synergy (resources in our separate operations can be combined to create a valuable new idea/practice/product)
from Burt, "The social capital of structural holes" (2002, The New Economic Sociology). The consequences of the
information diversity associated with network brokerage is productively elaborated at length in economist Scott
Page's 2007 book, The Difference: How the Power of Diversity Creates Better Groups, Firms, Schools and Societies.
Strategic Leadership
Network Brokerage (page 18)
FIRST CAUTION: Returns to network brokerage are a
probability, not a certainty. Access to structural holes
merely "increases the risk of productive accident."
Patent co-authoring
network from Lee Fleming & Matt Marx, "Managing creativity
from Brice Belisle, "Pet display clothing"
From Fleming & Marx, "Managing creativity in small worlds" (California Management Review, 2006)
in small worlds" (California Management Review, 2006; see Fleming et al. 2007 ASQ).
(US Patent 5,901,666 granted May 11, 1999).
Advantage from Varied Perspectives: Information arbitrage is
Strategic Leadership
Network Brokerage (page 19)
about framing as much as content. Problem vs. Paradox.* What point of
view, or frame of reference, will make my idea attractive?
The key is not to get "out of the box," so much as to see
from within a different box.* Failure here could be a good
idea over there.
Carl Segerstrom, in Chicago’s 2012 ADP, worked at
Pfizer when the Viagra trials were run. Carl sketched the
story: Trials showed that the new drug was a failure as a
heart medicine, so the trials were shut down and the test
samples were recalled. Subjects were asked to return
the test samples, and they usually do, but in this case, an
unusually high proportion of subjects did not return the
test samples. Someone asked, “let’s find out why they
aren’t returning the test samples,” which revealed the
profitable side-effect.
Originally, minoxidil was used
exclusively as an oral drug
(with the trade name ‘Loniten’)
to treat high blood pressure.
However, it was discovered to
have an interesting side effect:
hair growth. Minoxidil may
cause increased growth or
darkening of fine body hairs, or
in some cases, significant hair
growth. When the medication
is discontinued, the hair loss
will return to normal rate within
30 to 60 days.
*The "problem vs. paradox" point is nicely elaborated by David Doltish, Peter Cairo, and Cade Cowan in The Unfinished Leader (2014). The "out
of the box" point is nicely elaborated by Luc de Brabandere (2005), The Forgotten Half of Change: Achieving Greater Creativity through Changes
in Perception. See IDEO on the saying "fail often to succeed sooner," Stuart Firestein (2016) Failure, on the critical role failure plays in successful
science, and Ludwik Fleck (1979) Genesis and Development of a Scientific Fact, on the critical role that proto-ideas play in successful science.
In short, network brokerage is a process by which people clear sticky-information
markets. The rewards enjoyed by network brokers are compensation for clearing a
market that would otherwise not clear.
Therefore, variation between clusters/silos is essential to the value of brokerage. If
there are no information differences between social clusters, then there is no value
to moving information from one cluster to another.
Competition in open markets explicitly eliminates variation, though social clustering in networks usually
indicates variation in understanding and practice. For example, BP learning in the refining businesses.
Strategic Leadership
Network Brokerage (page 20)
Strong belief/culture/process/paradigm reinforce closed networks, and can obscure or blind people to
variation between subgroups within the network. For example:
— Pfizer drug trial protocol
— Talent out of context (able musician in D.C. metro train station)
— INSEAD student teams
— Coca Cola as a distribution company versus custodian of the Coca Cola brand
— "Hard" sciences & the negative correlation between age and contribution
look for use of right-wrong versus productive-unproductive or interesting-uninteresting
Personal experience is perhaps the most insidious blinder. Personal experience enriches understanding,
but it can also limit understanding. Many people are trapped in their limited personal experience. They
only hear/believe/understand knowledge consistent with what they’ve already experienced. The power of
understanding fundamental principles, and being able to re-frame problems in different ways, is that you
can reason your way through challenges that involve experiences you have not yet had — making you
valuable beyond whatever experience life has happened to give you personally.
Even within Frame,
Are You Thinking about the
Work Productively?
Modularity increases
the risk of productive accident.
Netscape’s Navigator was released under open-source license in March
1998 as Mozilla. It was re-designed for modularity to make it more
attractive to contributors. Networks below show module dependencies
before and after the re-design. ”Propagation cost” is the average
percentage of code that must be updated following a change in any one
module.
Mozilla version 1998-04-08
Strategic Leadership
Network Brokerage (page 21)
propagation cost:* 17.35%
Longitudinal Evolution
of Mozilla Propagation Cost*
Mozilla version 1998-12-11
propagation cost: 2.78%
From MacCormack, Rusnak,and Baldwin, “Exploring the structure of complex software designs” (2006, Management Science). For broad
discussion of modularity in high tech, see Baldwin and Clark, Design Rules: The Power of Modularity, 2000, MIT press.
And Social Context: Where did US time zones come from?
Until 1883 each United States railroad chose its own time
standards. The Pennsylvania Railroad used the "Allegheny
Time" system. By 1870 the Allegheny Time service extended
over 2,500 miles with 300 telegraph offices receiving time
signals. However, almost all railroads out of New York ran on
New York time, and railroads west from Chicago mostly used
Chicago time, but between Chicago and Pittsburgh/Buffalo
the norm was Columbus time, even on railroads which did
not run through Columbus. The Northern Pacific Railroad
had seven time zones between St. Paul and the 1883 west
end of the railroad at Wallula Junction.
Strategic Leadership
Network Brokerage (page 22)
In 1870 Charles F. Dowd proposed four time zones
based on the meridian through Washington, DC for North
American railroads. In 1872 he revised his proposal to
base it on the Greenwich meridian. Sandford Fleming,
a Canadian, proposed worldwide Standard Time at a
meeting of the Royal Canadian Institute on February 8,
1879. Cleveland Abbe advocated standard time to better
coordinate international weather observations and resultant
weather forecasts, which had been coordinated using local solar time. In 1879 he recommended
four time zones across the contiguous United States, based upon Greenwich Mean Time.
The General Time Convention (renamed the American Railway Association in 1891),
an organization of US railroads charged with coordinating schedules and operating standards,
became increasingly concerned that if the US government adopted a standard time scheme it
would be disadvantageous to its member railroads. William F. Allen, the Convention secretary,
argued that North American railroads should adopt a five-zone standard, similar to the one in use
today, to avoid government action. On October 11, 1883, the heads of the major railroads met in
Chicago at the Grand Pacific Hotel and agreed to adopt Allen's proposed system. ... Standard
time was not enacted into US law until the 1918 Standard Time Act.*
*Text comes from October 24, 2015 Wikipedia entry for "Standard time" (five zones include one east of Eastern zone). Map is
Dowd's 1884 fifth version advocating to railroaders the adoption of standard time zones. Engraving of William Allen is from Frank Leslie's Popular
Monthly (April 1884). For details on bureaucratic infighting over standard time, see Bartky, Selling the True Time (2000, Stanford University Press).
Illustration: Where did the M-16 come from?
Strategic Leadership
Network Brokerage (page 23)
Discussion Question*
Consequential ideas are typically attributed to special people, geniuses, in part to make us feel less
uncomfortable about our own ideas. True to form, an American armament expert describes Eugene
Stoner, the engineer who developed the M-16 assault rifle, as "an engineering genius of the first order."
Another describes him as "the most gifted small-arm designer since Browning." (Browning patented the
widely-adopted BAR and 45 automatic.)
1. Based on the brief history video, how would you describe Stoner's genius?
2. What circumstances might allow you or your colleagues to be as creative?
*Photos are from the video shown during the session. For discussion and references, see page 73 in Brokerage
and Closure. For sampling on the dependent variable, see Rosenzweig, “Misunderstanding the nature of
company performance: the halo effect and other business delusions,” 2007 California Management Review.
Average Z-Score Idea Value
Average Z-Score Idea Value
More Generally: Network Brokers Propose Better Ideas
A. Good ideas are associated with
many nonredundant contacts
2
(R = .89, n = 39, t = 12.05, -7.44)
B. Good ideas are associated
with large, open networks
(R2 = .64, n = 54, t = -9.67)
Strategic Leadership
Network Brokerage (page 24)
12+
NonRedundant Contacts
few ——— Structural Holes ——— many
Network Constraint
many ——— Structural Holes ——— few
Graphs show idea quality increasing with more access to structural holes in the networks around supply-chain managers in
a large electronics firm. Circles in graph A are average scores on the vertical axis within interval numbers of nonredundant
contacts. Circles in graph B are average scores on the vertical axis for five-point intervals of network constraint. Bold line is
the vertical axis predicted by a function of nonredundant contacts (graph A, linear and squared terms), or the natural logarithm
of network constraint (graph B). Association statistics in the graphs are computed from the displayed data.
see Figure 2.1 in Brokerage and Closure (or Figure 5 in Burt, "Structural holes and good ideas," 2004 American Journal of Sociology)
Brokerage, Good Ideas, and Innovation,
Digging a Little Deeper
J
3.5
0.9
E E
J
Y = a + b ln(C)
E
^
a
^b
t
Judge 1
6.42
-1.04
-5.8
Judge 2
4.08
-.63
-3.9
Combined
5.51
-.91
-7.4
E
E
G
2.5
G
0.8
^
P(no idea)
11.2 logit test statistic
0.7
J
J
0.6
J
E
J
E
J
0.5
E
G
J
G
G
E
10
20
30
40
50
G
C
C
G
G
C
G
60
C
C
J
G
G
1
J
C C
70
80
90
100
10
C
20
30
40
C
C
^
P(dismiss)
5.5 logit
C
test statistic
C
J
G
G
C C
J
G
G
0.4
J
E
E
1.5
J C
J
J
G
2
C
E
E
E
". . . for those ideas that were
either too local in nature,
incomprehensible, vague,
or too whiny, I didn't rate them"
50
60
70
80
90
Probability
Management Evaluation of Idea's Value
Strategic Leadership
Network Brokerage (page 25)
E
3
J
J
E
0.3
0.2
0.1
0
100
Network Constraint (C) on Manager Offering Idea
from Figure 2.1 in Brokerage and Closure (or Figure 5 in Burt, "Structural holes and good ideas," 2004 American Journal of Sociology)
In Sum, Brokers Do Better
B. High achievement is associated
with large, open networks
(R2 = .36, n = 85, t = -6.78)
(evaluation, compensation, promotion)
Z-Score Residual Achievement
A. High achievement is associated with
many nonredundant contacts
(R2 = .39, n = 125, t = 6.52, -4.12)
Strategic Leadership
Network Brokerage (page 26)
24+
NonRedundant Contacts
few ——— Structural Holes ——— many
Network Constraint
many ——— Structural Holes ——— few
Graphs show achievement — evaluation, compensation, and promotion — increasing with more access to structural holes
in six populations (analysts, bankers, and managers in Asia, Europe, and North America; Burt, 2010:26, cf. Burt, 2005:56).
Circles in graph A are average scores on the vertical axis within integer numbers of nonredundant contacts. Circles in graph
B are average scores on the vertical axis for five-point intervals of network constraint. Bold line is the vertical axis predicted
by a function of nonredundant contacts (graph A, linear and squared terms; dashed regression line is through zero to six
nonredundant contacts, R2 = .70, n = 39, t = 4.97, -2.70), or the natural logarithm of network constraint (graph B). Association
statistics in the graphs are computed from the displayed data.
Three Summary Points
Network Structure Is a Proxy for the Distribution of Information
For reasons of opportunity, shared interests, experience — simple inertia — organizations
and markets drift toward the bridge-and-cluster structure known as a “small world.” Over
time, information becomes "sticky" within clusters, different between clusters.
In Which Network Brokers Have a Competitive Advantage
Bridge relations across the structural holes between clusters provide information breadth,
timing, and arbitrage advantages, such that network brokers managing the bridges are at
higher risk of “productive accident” in detecting and developing good ideas. They are the
source of significant innovation in organizations and markets. In return, network brokers
tend to be better compensated than peers, more widely celebrated than peers, and
promoted more quickly to senior rank relative to peers; in short, brokers do better.
Strategic Leadership
Network Brokerage (page 27)
Three Points Follow from the Link between Network Brokerage & Innovation
- Closed networks do not identify unintelligent managers so much as expert specialists.
- Innovation is an import/export process. Value is not created at the innovation source. It
is created each time productive knowledge produces innovation in a target audience.
- Innovation depends on the network as well as the person. Innovation does not depend
on individual genius so much as it depends on employees finding opportunities to
broker knowledge from where it is routine to where it would create value.
Strategic Leadership
Network Brokerage (page 28)
Appendix
Materials
Appendix I:
Example
Network
Questionnaire
for a
Web Survey
Strategic Leadership
Network Brokerage (page 29)
for discussion
of these slides
and
how to collect
network data,
see Appendix A,
"Measuring the
Network,"
in
Neighbor Networks.
Figure A1 in Neighbor Networks
Strategic Leadership
Network Brokerage (page 30)
Appendix I,
continued
Figure A2 in Neighbor Networks
Appendix II: Network Metrics*
from Burt, "Formalizing the argument," (1992, Structural Holes); "Gender of social capital"
(1998, Rationality and Society); Appendix B "Measuring Access to Structural Holes," (2010, Neighbor Networks).
Network brokerage is typically measured in terms of opportunities to connect people. When everyone you know is connected
with one another, you have no opportunities to connect people. When you know a lot of people disconnected from one another,
then you have a lot of opportunities to connect people. “Opportunities” should be emphasized in these sentences. None of
the usual brokerage measures actually measures brokerage behavior. They index opportunities for brokerage. Reliability and
cost underlie the practice of measuring brokerage in terms of opportunities. It is difficult to know whether or not you acted on a
brokerage opportunity. One can know with more reliability whether or not you had an opportunity for brokerage. Acts of brokerage
could be studied with ethnographic data, but the needed depth of data would be expensive, if not impossible, to obtain by the
practical survey methods used to measure networks.
Strategic Leadership
Network Brokerage (page 31)
Good reasons notwithstanding, the practice of measuring brokerage by its opportunities rather than its occurrence means
that performance has uneven variance across levels of brokerage opportunities. Performance is typically low in the absence
of opportunities. Performance varies widely where there are many opportunities: (1) because some people with opportunities
do not act upon them and so show no performance benefit, (2) because it is not always valuable to move information between
disconnected people (e.g., explain to your grandmother the latest technology in your line of work), or (3) because the performance
benefit of brokerage can occur with just one key bridge relationship. A sociologist might do more creative work because of
working through an idea with a colleague from economics, but that does not mean that she would be three times more creative
if she also worked through the idea with a colleague from psychology, another from anthropology, and another from history. The
above three points can be true of brokerage measured in terms of action, but under the assumption that people invest less in
brokerage that adds no value, the three points are more obviously true of brokerage measured in terms of opportunities. It could
be argued that people more often involved in bridge relations are more likely to have one bridge that is valuable for brokerage,
and to understand how to use bridges to add value, but the point remains that the network measures discussed below index
opportunities for brokerage, not acts of brokerage.
Bridge Counts
Bridge counts are an intuitively appealing measure. The relation between two people is a bridge if there are no indirect connections
between the two people through mutual contacts. Associations with performance have been reported measuring brokerage
with a count of bridges (e.g., Burt, Hogarth, and Michaud, 2000:Appendix; Burt, 2002).
Constraint
I measure brokerage opportunities with a summary index, network constraint. As illustrated on the next page, network constraint
begins with the extent to which manager i’s network is directly or indirectly invested in the manager’s relationship with contact j
(Burt 1992: Chap. 2): cij = (pij + Σqpiqpqj)2, for q ≠ i,j, where pij is the proportion of i’s network time and energy invested in contact
See Appendix III to get free software to do these calculations for you. We use the software in the follow-on course, 39006.
Illustrative
Network and
Computation
A
B
F
Strategic Leadership
Network Brokerage (page 32)
Constraint
measures the
extent to which a
network doesn't
span structural
holes
C
E
D
Network constraint measures the extent to which your network time and energy
is concentrated in a single group. There are two components: (direct) a contact
consumes a large proportion of your network time and energy, and (indirect) a
contact controls other people who consume a large proportion of your network
time and energy. The proportion of i’s network time and energy allocated to j, pij,
is the ratio of zij to the sum of i’s relations, where zij is the strength of connection
between i and j, here simplified to zero versus one.
cij = (pij + Σq piqpqj)2 q ≠ i,j
network data
contact-specific
constraint (x100):
A
B
C
D
E
F
15.1
8.5
2.8
4.9
4.3
4.3
100(1/36)
A
B
C
D
E
F
gray dot
total 39.9 = aggregate constraint (C = Σj cij)
.
1
0
0
1
1
1
1
.
0
1
0
0
1
0
0
.
0
0
0
1
0
1
0
.
0
0
1
1
0
0
0
.
0
1
1
0
0
0
0
.
1
1
1
1
1
1
1
.
j, pij = zij / Σqziq, and variable zij measures the strength of connection between contacts i and j. Connection zij measures the lack
of a structural hole so it is made symmetric before computing pij in that a hole between i and j is unlikely to the extent that either
i or j feels that they spend a lot of time in the relationship (strength of connection “between” i and j versus strength of connection
“from” i to j; see Burt, 1992:51). The total in parentheses is the proportion of i’s relations that are directly or indirectly invested
in connection with contact j. The sum of squared proportions, Σjcij, is the network constraint index C. I multiply scores by 100
to discuss integer levels of constraint.
The network constraint index varies with three network dimensions: size, density, and hierarchy. Constraint on a person
is high if the person has few contacts (small network) and those contacts are strongly connected to one another, either directly
(as in a dense network), or through a central, mutual contact (as in a hierarchical network). The index, C, can be written as
the sum of three variables: Σj(pij)2 +2Σjpij(Σqpiqpqj) + Σj(Σqpiqpqj)2. The first term in the expression, C-size in Burt (1998), is a
Herfindahl index measuring the extent to which manager i’s relations are concentrated in a single contact. The second term,
C-density in Burt (1998), is an interaction between strong ties and density in the sense that it increases with the extent to which
manager i’s strongest relations are with contacts strongly tied to the other contacts. The third term, C-hierarchy in Burt (1998),
measures the extent to which manager i’s contacts concentrate their relations in one central contact. See Burt (1992:50ff.;
1998:Appendix) and Borgatti, Jones, and Everett (1998) for discussion of components in network constraint.
Strategic Leadership
Network Brokerage (page 33)
Size
Network size, N, is the number of contacts in a person's network. In graph-theory discussions, the size of the network around
a person is discussed as “degree.” For non-zero network size, other things equal, more contacts mean that a manager is more
likely to receive diverse bits of information from contacts and is more able to play their individual demands against one another.
Network constraint is lower in larger networks because the proportion of a manager’s network time and energy allocated to any
one contact (pij in the constraint equation) decreases on average as the number of contacts increases.
Density
Density is the average strength of connection between contacts: Σ zij / N*(N-1), where summation is across all contacts i and
j. Dense networks are more constraining since contacts are more connected (Σqpiqpqj in the constraint equation). Contact
connections increase the probability that the contacts know the same information and eliminate opportunities to broker information
between contacts. Thus, dense networks offer less of the information and control advantage associated with spanning structural
holes. Density is only one form of network closure, but it is a form often discussed as closure.
Hypothetical networks in the figure on page 35 illustrate how constraint varies with size, density, and hierarchy. Relations
are simplified to binary and symmetric in the networks. The graphs display relations between contacts. Relations with the person
at the center of the network are not presented (that person at the center is referenced by various labels such as "you," "ego,"
or "respondent"). The first column in the figure contains examples of sparse networks (zero density). No contact is connected
with other contacts. The third column of the figure contains maximum-density networks (density = 100). Every contact has a
strong connection with each other contact. At each network size, constraint is lower in the sparse-network column.
Hierarchy
Density is a form of closure in which contacts are equally connected. Hierarchy is another form of closure in which a minority of
contacts, typically one or two, stand above the others for being more the source of closure. The extreme is to have a network
organized around one contact. For people in job transition, such as M.B.A. students, that one contact is often the spouse. In
organizations, hierarchical networks are sometimes built around the boss.
Hierarchy and density both increase constraint, but in different ways. They enlarge the indirect connection component
in network constraint (Σqpiqpqj). Where network constraint measures the extent to which contacts are redundant, network
hierarchy measures the extent to which the redundancy can be traced to a single contact in the network. The central contact
in a hierarchical network gets the same information available to the manager and cannot be avoided in manager negotiations
with each other contact. More, the central contact can be played against the manager by third parties because information
available from the manager is equally available from the central contact since manager and central contact reach the same
people. Network constraint increases with both density and hierarchy, but density and hierarchy are empirically distinct measures
and fundamentally distinct with respect to social capital because it is hierarchy that measures social capital borrowed from a
sponsor.
Strategic Leadership
Network Brokerage (page 34)
To measure the extent to which the constraint on a person is concentrated in certain contacts, I use the Coleman-Theil
inequality index for its attractive qualities as a robust measure of hierarchy (Burt, 1992:70ff.). Applied to contact-specific constraint
scores, the index is the ratio of Σj rj ln(rj) divided by N ln(N), where N is number of contacts, rj is the ratio of contact-j constraint
over average constraint, cij/(C/N). The ratio equals zero if all contact-specific constraints equal the average, and approaches
1.0 to the extent that all constraint is from one contact. Again, I multiply scores by 100 and report integer values.
In the first and third columns on the next page, no one contact is more connected than others, so all of the hierarchy scores
are zero. Non-zero hierarchy scores occur in the middle column, where one central contact is connected to all others who are
otherwise disconnected from one another. Contact A poses more severe constraint than the others because network ties are
concentrated in A. The Coleman-Theil index increases with the number of people connected to the central contact. Hierarchy
is 7 for the three-contact hierarchical network, 25 for the five-contact network, and 50 for the ten-contact network. This feature
of hierarchy increasing with the number of people in the hierarchy turns out to be important for measuring the social capital of
outsiders because it measures the volume of social capital borrowed from a sponsor, which strengthens the association with
performance (this point is the focus of the later session on outsiders having to borrow network access from a strategic partner).
Note that constraint increases with hierarchy and density such that evidence of density correlated with performance can
be evidence of a hierarchy effect. Constraint is high in the dense and hierarchical three-contact networks (93 and 84 points
respectively). Constraint is 65 in the dense five-contact network, and 59 in the hierarchical network; even though density is
only 40 in the hierarchical network. In the ten-contact networks, constraint is lower in the dense network than the hierarchical
network (36 versus 41), and density is only 20 in the hierarchical network. Density and hierarchy are correlated, but distinct,
components in network constraint.
Partner
Networks
Clique
Networks
A
A
A
C
B
Strategic Leadership
Network Brokerage (page 35)
contacts
density x 100
hierarchy x 100
constraint x 100
from:
A
B
C
D
E
nonredundant contacts
betweenness (holes)
C
B
3
0
0
33
3
67
7
84
3
100
0
93
11
11
11
3.0
3.0
44
20
20
1.7
0.5
31
31
31
1.0
0.0
A
A
A
E
Larger
Networks
D
B
B
D
C
E
B
D
C
ok
B
C
5
0
0
20
5
40
25
59
5
100
0
65
4
4
4
4
4
5.0
10.0
36
6
6
6
6
3.4
3.0
13
13
13
13
13
1.0
0.0
10
0
0
10
10.0
45.0
10
20
50
41
8.2
18.0
10
100
0
36
1.0
0.0
Still Larger
Networks
contacts
density x 100
hierarchy x 100
constraint x 100
nonredundant contacts
betweenness (holes)
e rs
Network Density
E
D
Cliques
Br
contacts
density x 100
hierarchy x 100
constraint x 100
from:
A
B
C
nonredundant contacts
betweenness (holes)
C
Partners
Network Hierarchy
Small
Networks
Broker
Networks
To keep the diagrams simple, relations with ego are not presented.
Network Constraint
decreases with number of contacts
(size), increases with strength of
connections between contacts
(density), and increases with sharing
the network (hierarchy).
This is Figure 1 in Burt, "Reinforced Structural Holes,"
(2015, Social Networks, an elaboration of Figure B.2
in Neighbor Networks). Graph above plots density
and hierarchy for 1,989 networks observed in six
management populations (aggregated in Figure
2.4 in Neighbor Networks to illustrate returns to
brokerage). Dot-circles are executives (MD or more
in finance, VP or more otherwise). Hollow circles are
lower ranks. Executives have significantly larger,
less dense, and less hierarchical networks.
Network Constraint (x 100)
Strategic Leadership
Network Brokerage (page 36)
many ——— Structural Holes ——— few
Network Constraint (x 100)
many ——— Structural Holes ——— few
Ego-Network Betweenness
-.71 correlation
with log constraint
(Number Monopoly-Access Holes)
Ego-Network Betweenness
(Number Monopoly-Access Holes)
NonRedundant Contacts
-.90 correlation
with log constraint
R2 = .99
R2 = .92
NonRedundant Contacts
few ——— Structural Holes ——— many
The Network Measures of Access to Structural Holes
Are Strongly Correlated
These are network metrics for 801 senior people in two organizations analyzed in Burt, "Reinforced structural
holes" (2015, Social Networks). One organization is a center-periphery network of investment bankers (circles).
The other is a balkanized network of supply-chain managers in a large electronics company (squares). The
point is that networks rich in structural holes by one measure tend to be rich in the other measures.
Structural Folds Indicate Access to Structural Holes
11
7
3
14
1
4
2
6
5
Strategic Leadership
Network Brokerage (page 37)
Kind of Network
8
12
9
10
15
13
Network
Effective Size
Network
Size
(NonRedundant
Constraint
(Contacts)
Contacts)
Ego-Network
Betweenness
(Structural
Holes)
16
Reinforced Holes
(RSH)
Raw
Normalized
Ego-Network
Modularity
(Newman Q)
Closed (3, 4, 5, 7,
8, 9, 11, 12, 13,
14, 15, 16)
3
1.0
92.6
.00
.00
0%
.00
Broker (1)
2
2.0
50.0
1.00
.75
75%
.00
Broker (2, 6)
4
2.5
58.3
3.00
1.75
29%
.00
Fold Broker (10)
6
4.0
46.3
9.00
6.00
40%
.50
This is Figure 3 in Burt, "Reinforced structural holes" (2015, Social Networks),
based on the above networks in Figure 1 of Vedres and Stark, "Structural
folds: generative disruption in overlapping groups" (2010, American Journal of
Sociology). Correlations to the right are across the 801 bankers and managers
analyzed in the 2015 article.
Log Constraint
1.00
Effective Size
-.90
1.00
EN Betweenness
-.71
.88
1.00
RSH
-.71
.93
.91
Strategic Leadership
Network Brokerage (page 38)
*node data
id
ego
A
B
C
D
E
F
*tie data
from to tie
ego A 1
ego B 1
ego C 1
ego D 1
ego E 1
ego F 1
A B 1
E A 1
F A 1
B D 1
Appendix III: NetDraw Quick Start
Making your own sociograms and computing network metrics
for a group, project, organization, or market
1. DOWNLOAD THE FREE NETWORK SOFTWARE, NETDRAW (https://sites.google.com/site/netdrawsoftware/
download, then click on the "Exe only" option)
2. TYPE THE 21 LINES TO THE RIGHT INTO YOUR WORD PROCESSOR AND SAVE AS A TEXT FILE ENDING
IN .vna The 21 lines are a roster of people in the network followed by a roster of relations (e.g., ego has a relation to
person A at strength 1). These data define the network illustrating network constraint on page 32 in this handout.
3. LOAD THE .vna FILE INTO NETDRAW (“File” menu, “open” option, then “VNA text file” and “Complete”)
4. GENERATE A SPATIAL DISPLAY OF THE NETWORK (“Layout” menu, “graph theoretic layout” option, then
“spring embedding”). YOU SHOULD GET THE SOCIOGRAM BELOW. (See next page for basic nets discussed in
this course.)
Now play around to learn the wide capabilities of the software. Click and drag a node to move it and its relations
around. Remove arrows by clicking on the arrow button to the right of the row of command buttons just above the
sociogram display. Save the sociogram to a file for editing, pasting, and printing (“File” menu, “save diagram as”
option, then “metafile”).
D
For more complex edits, such as computing network metrics (“Analysis” menu,
“structural holes” option, then “ego network model” and save the data), see the
short user guide you can download on the page where you clicked "Exe only." If
you are not comfortable using new software, it might be wise to bring in someone
who can play with the software then brief you. FOR TEXT EXPLAINING THE
NETWORK METRICS, see Appendix II in this handout. Caution: Some versions
of NetDraw compute incorrect values of network constraint for isolates. Network
constraint is infinite for isolates, so constraint should be its maximum of one.
Some versions of NetDraw report a value of zero for infinity. This issue is not
likely an issue for you since you probably have contacts, else you wouldn't be
using the software.
E
ego
F
C
B
A
Appendix IV: Network Endogeneity
Most Distributed Independent Network Effect Most Centralized
Leadership
Leadership
(slow, happy)
(fast, unhappy)
A
C
C
B
D
B
B
E
B
C
D
C
D
A
A
E
Strategic Leadership
Network Brokerage (page 39)
CIRCLE (50.4 sec)
N
NC
Happy
A
2
50
58.0
B
2
50
C
2
D
E
CHAIN (53.2 sec)
N
NC
Happy
A
1
100
45.0
64.0
B
2
50
50
70.0
C
2
2
50
65.0
D
E
2
50
71.0
Avg
2.0
50.0
65.6
A
E
D
Y-NETWORK (35.0 sec)
N
NC
Happy
A
1
100
46.0
82.5
B
1
100
50
78.0
C
3
2
50
70.0
D
E
1
100
24.0
Avg
1.6
70.0
59.9
WHEEL (32.0 sec)
N
NC
Happy
A
1
100
37.5
49.0
B
1
100
20.0
33
95.0
C
4
25
97.0
2
50
71.0
D
1
100
25.0
E
1
100
31.0
E
1
100
42.5
Avg
1.6
76.7
58.4
Avg
1.6
85.0
44.4
The four networks are from the Bavelas-Leavitt experiments on leadership in task groups. The WHEEL is a traditional bureaucracy in
which C is in charge. The other three networks involve distributed leadership (all five people in the CIRCLE; B, C, and D in the CHAIN;
C and D in the Y-NETWORK). More distributed leadership is associated with more messages, slower task completion, and greater
enjoyment. Speed, messages, and enjoyment scores are from Leavitt (1951). Number of contacts (N) and network constraint (NC) are
computed from binary ties in the sociograms (number of contacts equals number of non-redundant contacts in these structures).
Figure 2.4 in Burt (2017, Structural Holes in Virtual Worlds)
Information
messages
Strategic Leadership
Network Brokerage (page 40)
Network Constraint
Mean Enjoyment Score
Mean Messages Sent
Answer
messages
Enjoyment
after first trial
Enjoyment
after last trial
Network Constraint
Times Cited as Group Leader
Behavioral and Opinion Correlates of Network Brokers
Network Constraint (
)
A. Network brokers tend to
distribute answers, people
in moderately constrained
positions tend to be conduits
for informational messages.
B. Network brokers are least
happy initially, but eventually
become the most pleased with
the experience.
C. The final outcome, by the
end of the experiment, is that
network brokers are most
likely to be recognized as the
unofficial group leader.
Data are from Leavitt (1949: Table 30,
following page 62).
Data are from Leavitt (1949:Table 29,
pages 60-61; "How did you like your job
in the group?).
Data are from Leavitt (1949: Table
8, page 38; “Did your group have a
leader? If so, who?”).
Figure 2.5 in Burt (2017, Structural Holes in Virtual Worlds)
30
XXX
X
X
X
X
XX
X
X X
X
XXX X X X
X
Appendix
V:
Strategic Leadership
Network Brokerage (page 41)
National
Differences
in
Business
Culture
Years Acquainted with Contact
X
25
20
X
XX X
X
X
X
X X XXXXX
X
X
X
X
X
X XXXX
X
XX
X
XXX
X XX
X
XX X X X
X X
XX
X X XX
X X
X XX
X
X
XXXXXX
XXX
XX
X X X
X
X
XX
X
X
X
X X
X
X XXX
X X
XX
X
X XXX
XX
X XXXX
X X
X
15
10
5
X
0
0
5
15
10
X XX
X
X
X
XX
XX
X X
X
XX
X
X
X
X
X XX X X X
XX
X
X
X
X
X
XX
X
X X
X
X
XX
X
XX
XXX
X XX
X
X
X
X
XX
X
XX
X
XX
XX
X
X
X
X
X
X
X X
X
X X
X
X
X XX X
X
X
X X
XX
X
XX XX
X
XXX
X
XX
XX
X
XX
X
X
X
X
X
XX
X
X
XX
X
X
X
X
X
X
XX
X
X X
X
X X
X
X
X
X
X XX
XX
X XX
XX
X
X
X
X
X
X
X
X
X
X
X
X X
XX
X
X
XX
XX
X
X
X
XX X
X
X XX
20
25
French Manager Years in the Firm
30
30
25
20
15
10
5
0
X
X
X
XX
XX X
X XX X
X
X
X X
X
X X
X
X
X
X
X
XXX
X
X
X
X
X
X
X
XX
XXX
XX
XX X X X
X
X
XX
XX
XX
XX
X X
X
X XX
X X
XX
X
X
X X
X
X
XX
X X
X XXX
X
XX X
X
XX
X
X
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
XX
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
XXX
XX
X
XX
X
X
X
X
X
XX
X X
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
XX X
X
X
X
X
X
X
XX
XX
X
X
X
X
XXXX
X
X
XX
X
XX X X
XX
X
X
X
XXXX
XX
X
X
X X X
XXX
X
X
X
X
X
XX
XX
X
X
XX
X
X
X
X
XX
X
XXXX
XX
X
X
X
XX
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
XX
X
XX
X
X
X
X
X
X
X
X
XX
X
X
X
X
XX
X
X
X
X
X
XX
X
X
X
X
XX
X
X
XXXX
X
X
X
X
XX
X
X
X
X
X
X
XX
X
X
X
X X
X
X
X
X
X
X
X
X
X
XX
X
X
X
XX X
X
X
X
X
X
X
X
X
X
X
X
X
X XX X X
X
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
XX
X
X
X
X
X
X
XX X X
XX
X X XX
X
X
X
X
X
X
X
X
X
XX
X
X
X
XX
X
XX
XXXX
XX X
XXX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
XX
X
X
X
XX
XX
X
X
X
X
X
X
X
X
XX
X X
XX
X
XXX
X
X
XX
X
X
X
X
X
X
XX
X
XXX
X
X
X
X
X
XX
X
X
X
X
X
XX
X
X X
XX
XX
X
XX
X
X
X
X
X
XX
X
X
X
XX
X X
X
XX
X
X XXX
X
X
X XX X
X
X
X
X
X
X XXX
X
X
X
X
XXX
XX
X
XXX X
X
X
X
XXX
XX
XXX
X
XX
X
X
X
XX
X
X X
XX
XX
X
X
XX
XX
X
X
X
X
X
X XX
X
X X X
X
X
X X
XX
X
X
XX
X
XX
X
X
X
X
XX
X
X X
X
X
X
XXX
XX
X
X
XX
X
X
XXX
X
X
X
X
XX
X
X
X
X
X
XX
X X
XX
X
X
X
XX
XX
XX
XX
X
X
X
X XXXX
X
X X
XXXX
X
XXX
X X
X
XX
X
X X X X X
0
5
15
10
20
25
American Manager Years in the Firm
Colleague Relations Predating Entry into the Firm
French Managers
American Managers
Years in
the Firm
Number
Colleagues
% Known
Before Firm
Mean Years
Known
Number
Colleagues
% Known
Before Firm
Mean Years
Known
0 to 10
105
26%
5.2
691
81%
12.6
11 to 20
160
15%
8.2
875
42%
13.5
Over 20
391
5%
10.3
129
6%
14.9
Total
656
11%
9.0
1695
55%
13.0
from Burt, Hogarth, and Michaud "The social capital of French and American managers" (2000, Organization Science)
30
Distinctions Between Kinds of Relations
(relations close together reach the same contacts)
supervisor
J
less
J close
knew
before J
valued
J
less than
monthly
close
weekly
J
monthly
less than
monthly
J
J
J
difficult
J
discuss
exit
J
Strategic Leadership
Network Brokerage (page 42)
3-9
American
Managers
other
J
J
subordinate
discuss
personal
J
daily
J
J
J
esp
close
valued
J
J
close
10+
J
esp
close
difficult
French
Managers
supervisor
J
buy-in
J
J
monthly
J
distant
JJ
socialize
1-2
less
close
J
3-9
J
J
J
distant
J
J other
subordinate
J 10+
J
buy-in
J
J
J
1-2
J
knew
before
weekly
J daily
J
discuss
personal
J
J
discuss
exit
J
socialize
from Figure 1.7 in Brokerage and Closure (cf. Figure A3 in Neighbor Networks)
Number of Groups
Average Character Level
60
50
All Characters
40
30
20
10
A
00 0
a
b
3
7
11
15
19
Probability of
Founding a
Group
3.0
23 25+
Strategic Leadership
Network Brokerage (page 43)
This is Figure 3.6 in Burt, Structural Holes in Virtual Worlds
(2017). Achievement increases with access to structural holes
between nonredundant contacts. Graph A describes EverQuest
II (EQ2). Graphs B and C describe achievement in Second Life.
a
Zero refers to isolates, people with no network as defined in text
and no group/guild affiliations.
b
Zero refers to people with no network, but affiliated with groups
in Second Life or guilds in EQ2.
Number of Groups
Founded
2.5
0.5
2.0
0.4
1.5
0.3
Founded Groups
Still Active
1.0
0.5
0.0
0.2
B
00 0
a
3
b
Effective Size
(Number NonRedundant Contacts)
Detail on Achievement and
Network in Second Life
0.6
7
11
15
19
0.1
0
23 25+
Number NonRedundant Friends
00a 0b
C
3
7
11
15
19
23 25+
90
80
Total Members across
All Groups Founded
70
60
50
40
Members in
Largest
Group Founded
30
20
10
0
Members in Groups Founded
Player’s Primary
Character
0.7
Probability of Founding
3.5
70
Strategic Leadership
Network Brokerage (page 44)
(for 77 electronic firms with R&D departments)
Predicted Probability of Company Patents
Returns to Brokerage for Entrepreneurs
in the Renewal of the Chinese Economy
Sample of 700 Chinese CEO
entrepreneurs in 2012. Networks
are key contacts during important events
in company history, most valuable current
contacts, most valuable senior employee,
and most difficult contact.
1%
r = .84
Network Constraint
many ——— Structural Holes ——— few
50%
mean network size
network density
non-redundant contacts
network constraint (x100)
4
11.0
1.0
27.2
mean percent family
frequency (days)
years known contacts
0.0
1.0
3.1
0.0
11.6
9.7
mean number of employees
year company founded
14
1985
67
2001
99%
6
10
46.7 100.0
3.5
7.6
51.2 81.8
426
60.0
CEO's
50.2
cite no
26.1
nuclear or
extended
1105
family!
2010
(61%)
NOTE — The above graph contains the 77 electronics companies with R&D departments. The Y axis is from a logit equation predicting whether or not each
of the 700 companies had filed for patents (73% had not filed for patents). The predictors include years since founding (0.79 z-score), number of company
employees (2.35 z-score), whether the company had an R&D department (9.19 z-score), log network constraint (-2.97 z-score, P < .01), an adjustment
for lower patenting in textiles, transportation equipment, and medicine manufacturing (-2.17 z-score), and an adjustment for a weaker network effect in the
lower patenting industries (2.16 z-score).
For contextual background on the sample CEOs, see Nee and Opper, Capitalism from Below: Markets and Institutional Change
in China (2012, Harvard University Press). Also see Merluzzi (2013, "Social Capital in Asia," Social Science Research).