 

m“ 1» -~v
q n. Eu '9’,

‘ %

v.47 '63::
‘ Hr ..

was

111‘:

. _.
N

gaff?" '«jr inking

ﬁnik

[7
u

. ”km. ‘

.
wamgn , _

‘7 i'{":

 

 

 

 

“ a
gut v' ‘ .
‘ uqﬁpggjﬁ.
“2% “@333
59.. = -

rt

{ 'v :
3v
tww.‘ a
»

.. u .r
(a, J.
T m
a”
._ vr-
4’ﬁécv

W

“3:03
. . $.59??th
ﬁber“ '

 

'2. LIBRARY
ZVOC Michigan State
University

This is to certify that the
dissertation entitled

THE INFLUENCE OF EXPERIENCE ON ORGANIZATIONAL
SEARCH, KNOWLEDGE CREATION, AND PERFORMANCE

presented by

MICHAEL J. MANNOR II

has been accepted towards fulfillment
of the requirements for the

PhD. degree in Management

 

 

zwﬂM/z

Major Professor’s SignatJre

06/24/2008

 

Date

MSU is an affirmative-action, equal-opportunity employer

-.—-.---------0-c-----o-------------o---—o-.—-—-u_u-.-.-.-.-.-.--.--.--.-.-.-.-.-

 

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K./ProilAcc&PresICIRC/DateDue‘indd

 

THE INFLUENCE OF EXPERIENCE ON ORGANIZATIONAL SEARCH,
KNOWLEDGE CREATION, AND PERFORMANCE
By

Michael J. Mannor II

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY

Department of Management

2008

ABSTRACT

THE INFLUENCE OF EXPERIENCE ON ORGANIZATIONAL SEARCH,
KNOWLEDGE CREATION, AND PERFORMANCE

By

Michael J. Mannor II

In this dissertation I examine the overall question of how organizations can learn from
different types of experience to create new knowledge and achieve superior performance.
I take a three essay approach to examining these issues. I begin by developing a new
theoretical approach to looking at organizational experience by drawing on theory from
micro-level research to explore how quantitative depth and qualitative breadth in
experience can differentially inﬂuence search and innovation in firms. In the next essay, I
move on to apply these ideas to knowledge creation in the pharmaceutical industry. These
ideas are examined empirically in 20 years of US. patent data. The final essay then
concludes by investigating the inﬂuence of these different types of experience on ﬁrm

performance among production companies in the US. film industry.

DEDICATION

I dedicate this work to the memory of my father and brother, both of whom enriched my
life tremendously and continue to be missed deeply.

iii

ACKNOWLEDGEMENTS

I owe a great deal of thanks to my beautiful wife Katie, my kids, my mom, and the
inspiring community of scholars at MSU. I would like to particularly thank the members
of my committee, Kent Miller, Gerry McNamara, Jamal Shamsie, and Don Colon. Also,
thanks to Sue Polhamus for all of her help throughout my doctoral program and with the
ﬁnal preparation of this document. The management department at Michigan State is a
great group of friends and colleagues, and I could imagine no better place to go to work

everyday for the last ﬁve years.

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................ vii
LIST OF FIGURES ......................................................................................................... viii
INTRODUCTION ............................................................................................................... 1
CHAPTER 1: CREATING BREAKTHROUGHS: EXAMINING THE
EXPERIENTIAL ROOTS OF SEARCH AND KNOWLEDGE CREATION ................... 6
Introduction .......................................................................................................................... 6
Conceptual Development ................................................................................................... 10
The Rich History of the Experiential Learning Tradition .............................................. 10
Shortcomings in the Historical Research on Experiential Learning .............................. 12
Advances toward a Broader Conceptualization of Experience ...................................... 14
Extending Traditional Models of Organizational Experience ........................................... 16
Extending Traditional Models of Experience Type ....................................................... 16
Extending Traditional Models of Experience Level ...................................................... 19
A New Multi-Level Model of Organizational Experience ............................................ 21
Applying This New Model of Experience To Search and Learning ................................. 22
The Mechanisms of Organizational Learning ................................................................ 22
The Critical Role of Search Routines in the Pursuit of Radical Innovation .................. 25
Supporting and Conflicting Norms ................................................................................ 27
Theory and Propositions .................................................................................................... 28
A New Model of Experience, Search, and Knowledge Creation .................................. 28
Individual Experience and Organizational Search Routines ......................................... 31
Organizational Experience and Organizational Search Routines .................................. 33
Organizational Experience and Organizational Norms ................................................. 36
The Moderating Inﬂuence of Organizational Norms ..................................................... 39
Connecting Search Routines to Innovation .................................................................... 42
Discussion and Conclusions .............................................................................................. 45
Implications of this Research for Organizational Learning Theory and Practice .......... 46
Future Directions ........................................................................................................... 48

CHAPTER 2: EXPERIENTIAL LEARNING AND KNOWLEDGE CREATION IN

THE PHARMACEUTICAL INDUSTRY ......................................................................... 50
Introduction ........................................................................................................................ 5 1
Theory and Hypotheses ...................................................................................................... 53
Experience and Knowledge Creation ............................................................................. 53
Experience Depth and Knowledge Creation .................................................................. 55

Experience Breadth and Knowledge Creation ............................................................... 57

Experiential Diversity and Knowledge Creation ........................................................... 60
Methods .............................................................................................................................. 65
Data and Setting ............................................................................................................. 65
Measures ........................................................................................................................ 66
Independent Variables ................................................................................................... 67
Dependent Variables ...................................................................................................... 71
Control Variables ........................................................................................................... 72
Econometric Analyses ................................................................................................... 73
Results ............................................................................................................................ 81
Discussion .......................................................................................................................... 83

CHAPTER 3: CREATING BLOCKBUSTERS: UNIVERSAL AND CONTINGENCY
PERSPECTIVES ON EXPERIENTIAL LEARNING AND PERFORMANCE IN

HOLLYWOOD .................................................................................................................. 89
Introduction ........................................................................................................................ 90
Theory and Hypotheses ...................................................................................................... 93

Universal and Contingency Perspectives on Experiential Learning and Performance..93
All Experience is Good Experience? The Universal Beneﬁts of Experience for

Performance ................................................................................................................... 95
A Contingent Perspective on Experience: Tradeoffs for Blockbuster Performance ..... 97
Experience Breadth and Blockbuster Performance ..................................................... 100
Experience Depth and Blockbuster Performance ........................................................ 102
Methods ............................................................................................................................ 103
Data and Setting ........................................................................................................... 103
Independent Variables ................................................................................................. 105
Dependent Variables .................................................................................................... 107
Controls ........................................................................................................................ 110
Analyses ....................................................................................................................... 1 11
Results .......................................................................................................................... 116
Discussion and Conclusion .............................................................................................. 117
CONCLUSION ................................................................................................................ 122
APPENDIX ...................................................................................................................... 127
REFERENCES ................................................................................................................ 130

vi

LIST OF TABLES

Table 1 — Different Types of Organizational Experience ...................................... 17
Table 2 — A New Framework of Organizational Experience .................................. 21
Table 3 — Correlations between Study Variables ................................................ 76

Table 4 — Fixed Effect Negative Binomial Regression Results (Knowledge Creation

Rate).. ................................................................................................................ 77
Table 5 — Zero-Inﬂated Negative Binomial Regression Results (Knowledge Creation

Impact 1) ................................................................................. 78
Table 6 — Zero-Inﬂated Negative Binomial Regression Results (Knowledge Creation

Impact 2) .................................................................................. 79
Table 7 — Regression Results Summary .......................................................... 80
Table 8 — Descriptive Statistics and Correlations ............................................. 112

Table 9 — Panel Regressions for Depth and Breadth of Experience to Overall
Performance ..................................................................................................... 113

Table 10 — Logistic Panel Regressions for Depth and Breadth of Experience to
Blockbuster Box Ofﬁce Performance .............................................. 114

Table 11 — Logistic Panel Regressions for Depth and Breadth of Experience to
Blockbuster Critical Ratings Performance ......................................... 115

vii

LIST OF FIGURES

Figure 1 - Full Conceptual Model of Experience, Search, and New Knowledge
Creation ................................................................................... 30

Figure 2 — Lorenz Curve for Box Office Revenue in Hollywood 99

viii

INTRODUCTION

The idea that organizations learn from experience has been well accepted in the
study of organizations for at least 80 years (Rohrback, 1927; Wright, 1936; Yelle, 1979).
In the course of this study the beneﬁts of experiential learning have been shown to be
important for the study of economic advantages in airplane manufacture (Alchian, 1962),
nuclear power plant construction (Zimmerman, 1982), truck production (Epple, Argote,
& Devadas, 1996), pizza-making (Darr, Argote, & Epple, 1995) and many other
industries. In recent years this line of research has branched out to study the strategic
advantages of experiential learning for the management of alliances (Hoang &
Rothaermel, 2005), the success of new entries (King & Tucci, 2002), the performance of
acquisitions (Zollo & Singh, 2004), and research and development investments (Nerkar
& Roberts, 2004). This rich history has launched books (Argyris & Schon, 1978),
corporate initiatives (McGregor, 2008), and the careers of many consultants.

However, in this dissertation it is argued that the conceptualization of experience
in this research tradition has been, collectively, rather narrow. Although the industry
settings have broadened, new methods have been applied, and new strategic outcomes
have been considered, a consistent theme throughout the literature has been the
conceptualization of experience as a quantitative assessment of cumulative experience
collected by an organization over time. Whether this is measured with the number of
years of exposure or the cumulative number of units produced (Argote, 1999), the
number of airplanes produced (Asher, 1961) or the number of pizzas made (Darr, et. al,
1995), research has consistently failed to recognize a wider perspective of the different

types or challenges presented by a ﬁrm’s history of experiences. By developing new

theory to better capture the inﬂuence of different types of experience, it may be possible
to re-examine existing ﬁndings, question established truths, and uncover new avenues for
research. Interestingly, this signiﬁcant oversight has occurred despite the encouragement
of various literature reviewers through the years to explore these broader experiential
perspectives (Levitt & March, 1988; Huber, 1991; Bapuji & Crossan, 2004).

Responding to this shortcoming, this dissertation begins with the development
and application of a new and more complex model of organizational experience. Building
on advances in the micro-level study of work experience and job performance
(Quinnones, Ford, & Teachout, 1995; Tesluk & Jacobs, 1998), this dissertation integrates
these ideas into a new model of organizational experience and then explores how this
new model can help to inform the study of search and innovation in organizations. In
subsequent chapters, this new approach to organizational experience is further extended
by investigating the inﬂuence of different types of experience on knowledge creation and
performance. Further, in applying these new perspectives of experience, there is a
consistent focus throughout the dissertation on the inﬂuences of different types of
experience on creating ‘breakthroughs’. This focus on breakthroughs is motivated in part
due to the fact that radical innovations have the potential to help organizations to achieve
ﬁrst-mover advantages (Lieberman & Montgomery, 1987), create barriers to entry
(Porter, 1980), better meet consumer demands (Adner & Levinthal, 2001), and generally
push the envelope of competitive pressure on their rivals (Utterback & Suarez, 1993).
Further, radical innovations have been shown to be responsible for between 48-93% of
the ﬁnancial payoffs from the patenting activities of ﬁrms (Scherer & Haroff, 2000).

Despite these substantial potential beneﬁts, academic research has been slow to explore

empirically the processes of radical innovation (Garcia & Calantone, 2002; Fleming,
2002), and organizations vary signiﬁcantly in their ability to create breakthrough
innovations (Ettlie, Bridges, & O’Keefe, 1984; Ahuja & Lampert, 2001). By applying
this new model of experience to the study of how organizations learn to create such
innovations, this work has the potential to open new doors and push into new
understanding of this important phenomenon.

In fact, despite the struggles of many ﬁrms, some organizations appear to have the
“winning formula” for consistently churning out great industry-changing products. In the
past luminaries such as Sony, IBM, GE, 3M, Apple, and Pﬁzer have served as models to
their peers and have been rewarded by investors for consistently churning out
revolutionary products. Although these models have been studied by competitors, few
have been able to create innovative environments as successful as these ﬁrms. Further,
many of these successful organizations have had difﬁculty themselves sustaining their
own successful innovation. Causal ambiguity and social complexity often make the
successful “formula” difficult to understand, even from the inside (Fleming, 2002).
Perhaps due to these difﬁculties, several of these luminary model organizations have
stumbled in recent years.

Thus, with the goal of developing a more complete perspective of experience in
organizations and then applying these new experiential ideas to the study of search,
knowledge creation, and performance, the structure of this dissertation is simple. First,
Chapter 1 begins by reviewing the rich history of research on experiential learning and
then offering a formal critique of how this research has been held back by narrow

conceptualizations of experience. Chapter 1 then introduces more complex ideas of

experience from the work experience and job performance literature and develops a new
model of the qualitative-diversity as well as the quantitative-intensity of experiences at
both individual and organizational levels of analysis. Chapter 1 continues by connecting
these new ideas of experience to the mechanisms of learning in organizations
(organizational norms and routines) to explore how these mechanisms are inﬂuenced by
these different types and levels of experiences. Chapter 1 concludes with an analysis of
how speciﬁc norms and routines inﬂuenced by these experiences contribute to the
development of radical innovations in organizations.

Chapters 2 and 3 build from this conceptual base to examine these ideas
empirically in two data settings — the pharmaceutical industry and Hollywood. Chapter 2
explores the inﬂuence of different types of organizational experience on knowledge
creation in the pharmaceutical industry. In this chapter I explore how different types of
organizational experience (experiential depth, breadth, and diversity) inﬂuence the rate
and impact of knowledge creation in the pharmaceutical industry. In chapter 3 the ideas
from chapter 1 are extended to the study of ﬁrm performance in Hollywood. Specifically,
the relation between different types of organizational experience and performance is
investigated. Building on experiential learning theory, two perspectives are advanced.
First, a universalistic approach is developed from the traditional learning literature,
building on the logic that all experience is good experience. Second, a more contingency-
based approach is hypothesized, in which the inﬂuence of different types of experience
on performance can vary signiﬁcantly, and even be negative in some cases.

In sum, these three chapters advance new perspectives of organizational

experience theoretically and empirically to advance experiential learning theory. Each

chapter is designed to be both integrated into the core themes of the dissertation, but also

with a clear modularly for future publication in a more traditional journal format.

CHAPTER 1

CREATING BREAKTHROUGHS: EXAMINING THE EXPERIENTIAL
ROOTS OF SEARCH AND KNOWLEDGE CREATION

OVERVIEW

This chapter extends theory on the experiential roots of organizational learning by
examining the different types and levels of such learning. This is achieved in two parts.
First, research from inside and outside organizational learning are brought together to
inform a new multi-Ievel framework of organizational experience. Second, these new
ideas are used to examine the experiential roots of norms and routines in organizations,

focusing on the context of search and knowledge creation in organizations.

The idea that organizations learn from experience has been well accepted in the
study of organizations for at least 80 years (Rohrback, 1927; Wright, 1936; Yelle, 1979).
In the course of this research the beneﬁts of experiential learning have been shown to be
important for the study of economic advantages in airplane manufacture (Alchian, 1962),
nuclear power plant construction (Zimmerman, 1982), truck production (Epple, Argote,
& Devadas, 1996), pizza-making (Darr, Argote, & Epple, 1995) and many other
industries. In recent years this research has branched out to study the strategic advantages
of experiential learning for the management of alliances (Hoang & Rothaermel, 2005),
the success of new entries (King & Tucci, 2002), the performance of acquisitions
(Haleblian & Finkelstein, 1999; Haleblian, Kim, & Rajagopalan, 2006; 20110 & Singh,
2004), and research and development investments (Sampson, 2005, Nerkar & Roberts,
2004). This rich history has launched books (Argyris & Schon, 1978), corporate
initiatives (Jana, 2007), and the careers of many consultants.

However, in this study I argue that the conceptualization of experience in this
research tradition has been, collectively, rather narrow. Although the industry settings
have broadened, new methods have been applied, and new strategic outcomes have been
considered, a consistent theme throughout the literature has been the conceptualization of
experience as a quantitative assessment of cumulative experience collected by an
organization over time. Whether this is measured with the number of years of exposure or
the cumulative number of widgets produced (Argote, 1999), the number of airplanes
produced (Asher, 1961) or the number of pizzas made (Darr, et. al, 1995), researchers
have consistently failed to recognize a wider perspective of the different types or

challenges presented by a ﬁrm’s history of experiences. By developing new theory to

better capture the inﬂuence of experience diversity, in addition to experience intensity, it
may be possible to re-examine existing findings, question established truths, and uncover
new avenues for research. Interestingly, this significant oversight has occurred despite the
encouragement of literature reviewers through the years to explore these broader
experiential perspectives (Levitt & March, 1988; Huber, 1991; Bapuji & Crossan, 2004).
Despite these cautions, a critical question that the current study seeks to answer is
why a historically limited theoretical conceptualization of organizational experience is
inadequate. In fact, I argue that several shortcomings leave existing models of
experiential organizational learning simplistic and insufﬁcient for advancing our
understanding of how organizations learn and create knowledge. First, simplistic models
of experiences are inadequate for understanding how diverse organizational norms and
routines are developed in organizations. Focusing on only the intensity of experiences
significantly biases any models of norms and routines to focus toward efﬁciency, missing
the entire domain of developing broad learning organizations. This is particularly
important, as norms and routines act to both protect and sustain business advantages from
competitors over time (Nelson & Winter, 1982), and are significantly inﬂuenced both by
organizational and individual experiences (Bettenhausen & Mumighan, 1985; Levitt &
March, 1988). Second, as existing models of experiential organizational learning have
both simplified the treatment of experience and consistently failed to recognize a multi-
level perspective (Huber, 1991; Bapuji & Crossan, 2004), understanding the processes
through which different types of experience interact across levels of analysis is
impossible. In fact, a wealth of research has been done at the individual level of analysis

that is almost completely missing from macro treatments of experience. This is evident,

in particular, in the fact that significant advances have been made in the study of
experience in industrial-organizational psychology (e. g. Quinnones, Ford, & Teachout,
1995; T esluk & Jacobs, 1998), but macro-level research has yet to embrace or integrate
such advances into organizational level models. This instance of research myopia is
unfortunate, as insights at the individual level have important implications at the
organizational level as well. Further, keeping our conceptualization of organizational
experience simple in terms of level and type has sheltered macro-level research from the
interesting but complicated issues of ﬁt and misﬁt between individual and organizational
experiences. Although these issues are complex, such situations are likely to have an
important inﬂuence on organizational processes, inﬂuences that are not able to be
captured by simple existing models of organizational experience.

Taking these ideas together, the bottom line is that it is virtually impossible to
theoretically disentangle the nature of experiential organizational learning with our
current conceptualizations of experience. To impact actual practice, we need new theory
that moves away from traditional assumptions and myopias to better encompass the
complex reality of the phenomenon, in which experiences are not all equal, and have very
different inﬂuences on the development of norms and routines.

Directly responding to these shortcomings, this study focuses on the development
and application of a new and more complex model of organizational experience. Toward
this end, this study begins with a critique of the traditional treatment of experience in the
organizational learning literature, and how several recent advances both inside and
outside the organizational learning literature can inform the development of a new model

of organizational experience. In particular, I speciﬁcally build on advances in the micro-

level study of work experience and job performance (Quinnones, Ford, & Teachout,
1995; Tesluk & Jacobs, 1998) to first integrate these ideas into a new multi—level model
of organizational experience. With this new experience model, I then move on to apply
the new model to understand how this more complex approach can help to inform the
study of norms and routines in organizations, and ultimately how organizations create
radical new knowledge. I conclude with a discussion of the implications of this research
for theory and practice, and directions for future research.
CONCEPTUAL DEVELOPMENT

The Rich History of the Experiential Learning Tradition

The ﬁrst step toward reaching the goal of understanding the inﬂuence of
experience on the development of radical innovations in organizations is to review the
long history of experiential learning research through the years. Few literatures in the
study of organizations have such an extensive and inter-disciplinary history as that of the
organizational learning from experience literature. Early work in this tradition introduced
the notion of learning from experience primarily in the manufacture of airplanes in the
1920’s and 1930’s (Rohrbach, 1927; Wright, 1936). Variously called “Ieaming curves”,
“progress functions”, or “experience curves” all of this research was directed at the
general proposition that cost and labor savings accrue with increasing levels of
experience. This research continued through World War II and beyond with
governmental assistance and aided by the support of the RAND corporation (Asher,
1961; Alchian, 1963; Garg & Milliman, 1961) focusing on the nature of this relation
between experience and savings in airplane manufacture. Building on this research,

others began examining this new leaming-by-doing perspective in the manufacture of

10

radar equipment (Preston & Keachie, 1964), rayon fiber (Hollander, 1965), production
systems (Baloff, 1966; 1971), and various other manufacturing contexts (Conway &
Shultz, 1959; Nadler & Smith, 1963; Hirsch, 1952). In a very interesting narrative review
of this research, Dutton, Thomas, and Butler (1984) explain that postwar empirical
studies continued to incorporate the progress function concept into economic theory by
studying the learning relationships in a wide variety of processes. The results of this
research clearly established that signiﬁcant improvements were associated with
experience but that there was considerable variation between processes, products, ﬁrms,
and facilities (Yelle, 1979; Dutton, et. a1, 1984).

Continuing this tradition, research carried on in great abundance in the 1980’s and
1990’s to build on this foundation but began to expand to a broader focus in terms of the
contexts of learning and the use of some new methods to assess learning from experience.
Still, the focus of this research remained the economic benefits of learning for cost
functions and reducing labor expense. Highlighting some of this research, strong learning
from experience effects were found in the construction of nuclear power plants
(Zimmerman, 1982), shipbuilding (Argote, Beckman, & Epple, 1990), semi-conductors
(Dick, 1991), biomedical devices (Ven de Ven & Polley, 1992), truck manufacturing
(Epple, Argote, & Devadas, 1996), and pizza-making (Darr, Argote, Epple, 1995). New
research methodologies employed in this research also began to bring insights from
simulation models to bear on experiential learning with interesting research from Muth
(1986), Hatch and Mowry (1998), and Gavetti and Levinthal (2000).

In addition to a broader contextual focus, in recent years research has also begun

to further expand the boundaries to explore how learning can impact strategic in addition

11

to economic outcomes. The question has turned from whether or not learning occurs to
what exactly organizations learn to do better. Clearly, organizations learn to reduce unit
and labor costs, but they also learn from their experiences to better understand relations
with more speciﬁc strategic consequences. Research in this tradition has examined the
effects of experience on strategic change (Amburgey, Kelly, Barnett, 1993), joint venture
success (Barkema, Shenkar, Verrneulen, & Bell, 1997), hotel survival (Baum & Ingram,
1998), new product introductions (Martin & Mitchell, 1998), the timing of new entry
(Martin, Saminathan, & Mitchell, 1998), the decision of where to locate a new business
(Baum, Li, & Usher, 2000), the success of new entries (King & Tucci, 2002), R&D
performance (Hoang & Rothaermel, 2005), acquisition performance (Haleblian &
Finkelstein, 1999), and innovation (Nerkar & Roberts, 2004).

Shortcomings in the Historical Research on Experiential Learning

Despite the strong history and tremendous breadth of the experiential learning
literature, it is the thesis of this study that this research has, collectively, taken a rather
narrow view of experience. Further, I argue that by taking a broader approach to the
study of experiential learning, the elimination of this historical restriction can lead to
powerful new insights. In making this claim it is ﬁrst important to understand the nature
of this historically narrow focus.

From its inception as a field of study, the leaming-by-doing literature has focused
most attention on just one type of experience — a simple quantitative assessment of the
time or amount of exposure an individual, organization, or population has had in relation
to a speciﬁc stimulus. For example. in the aircraft manufacture tradition, it was exposure

over time to an increasing number of aircraft building projects (Wright. 1936). In other

12

cases it was the number of nuclear plants constructed (Zimmerman, 1982), the number of
angioplasty surgeries a surgeon has completed (Kelsey, et. al, 1984), or the number of
pizzas a franchisee had produced (Darr, et. a1, 1995). In this view, it is the ability of such
individuals, organizations, and populations to observe this stimulus repeatedly over time,
reﬂect, make generalizations, and then adjust their behavior that is called ‘leaming’
(Kolb, 1984). It is through this type of quantitatively-intense learning that the benefits of
cost savings have been found to occur (Yelle, 1979). In more recent research new
strategic outcomes and methodologies have been employed, but the one variable that has
remained relatively consistent throughout the years is that experience has been
conceptualized using this narrow quantitative assessment of exposure to a stimulus.
Interestingly, the idea that narrow conceptions of experience need to be expanded
has emerged several times through the years in reviews of the organizational learning
literature. For example, in Levitt and March’s (1988) review, the authors point out that in
addition to the direct effects of cumulative experience totals, the complexity of the
experience may also be an important consideration. In another example, Huber (1991)
goes out of his way to make clear that “it is important to challenge narrow concepts of
organizational learning.” In regard to a more complex view of experience, he goes on to
suggest that the variety, not just the quantity, of experiences is important for
organizational learning, and speciﬁcally outlines methods through which organizations
can gain a wider set of experiences (Huber, 1991). In the most recent review, Bapuji and
Crossan make the point very succinctly, arguing “there is a need to know what type of

previous experience is useful for which situations” (2004: 403).

13

Advances toward a Broader C onceptualization of Experience

Responding to the need for research that broadens our conceptualization of
experience-based learning, two streams of research have emerged - one inside the ﬁeld of
organizational learning and one outside in the discipline of industrial—organizational
psychology.

From inside organizational learning research a small but growing stream of work
has begun to embrace a broader conceptualization of experience, though somewhat
inconsistently. For example, after noting that “many organizational learning studies have
an implicit assumption that the learning rate is maximized through specialization,”
Schilling, Vidal, Ployhart, and Marangoni (2003) look at how people can improve their
learning rates by doing ‘something else.’ Although the center of their research was still
focused on the rate of learning, the typical learning curve outcome, their lab study sparks
an interesting question — how is experience doing something ‘different’ unlike experience
doing more of the same? This line of reasoning builds on March’s (1991) characterization
of exploration versus exploitation motivations in organizational learning. March argued
that ﬁrms must make tough decisions in allocating scarce resources to manage the tension
between these two different paths, one of which leads to narrow experience and learning
(exploitation) and the other leads to broad experience and learning (exploration).
Building on these theoretical ideas, several recent empirical ﬁndings provide further
support for the argument that a more complex view of experience is needed in the
organizational learning literature. For example, in their study of the disk drive industry,
King and Tucci (2002) argue that at least two types of experience exist — ‘static’ and

‘transformational’ experience. Static experience is argued to be experience in existing

14

markets, but transformational experience is argued to be experience with ‘major
changes.’ Building on this distinction, their results suggest that these different categories
of experience have different effects on the strategic decisions of ﬁrms to enter new
markets. In another example, Nerkar and Roberts (2004) argue that important differences
exist between ‘proximal,’ ‘distal,’ and ‘general’ experiences. In their study of the
pharmaceutical industry, proximal experiences are those that are in the same technical
area as current research endeavors, distal experiences are experiences in other therapeutic
areas, and general experiences are those outside of therapeutic technologies but still in the
pharmaceutical domain. Their results again support the idea that differences in experience
lead to differences in strategic outcomes. In several additional empirical examples,
Haleblian and Finkelstein (1999), Haleblian, et a1. (2006), and Hoang and Rothaermel
(2005) all ﬁnd that different types of acquisition and alliance experience have different
effects on future performance. In all of these examples we see different but connected
logics that come to the same conclusion — not all experiences are the same. Building on
this simple but powerful realization, each suggests that a new theoretical foundation is
needed to push forward the experiential learning literature.

However, in addition to the few but growing number of promising new ideas
inside the organizational learning literature, even more significant advances have been
made outside the study of organizational learning. The common theme in this outside
work is that a more complex conceptualization of experience can bring new insight to
previously well-understood relationships. Specifically, in the field of industrial-
organizational psychology, research on the determinants of job performance in

organizations has begun to theoretically break apart experience into a much broader and

15

more complex set of related constructs. Two theoretical advances, one by Quinnones,
Ford, and Teachout ( 1995) and another by Tesluk and Jacobs (1998) have suggested that
we look at experience not only in quantitative terms, but in a much wider sense where
experience can be viewed as qualitatively diverse as well as quantitatively intense.
Arguing, similar to the claims made here, that the historical measurement of experience
in the job performance and compensation domains has been too simple and too narrow,
these authors make a compelling case for the more complex view of experience that they
develop (Tesluk & Jacobs, 1998). Building on these arguments, these authors propose a
new model of experience and suggest powerful new avenues for research that build from
this new model to re-analyze and extend the ﬁndings on experience and job performance
developed through the years.

Thus, the goal of the current research is to bridge the gap between the work being
done inside the organizational learning literature and the very promising work being done
in industrial-organizational psychology to provide a single integrated foundation for
moving forward the study of experience-based organizational learning. Although the
work in each domain has made important strides, it is by bringing these streams together
that we can work together toward a unified and conceptually enriched platform for future
research on experience-based organizational learning. This is the motivation for the next
section of this paper.

EXTENDING TRADITIONAL MODELS OF ORGANIZATIONAL EXPERIENCE
Extending Traditional Models of Experience Type
Building on these ideas, the ﬁrst distinction between experiences is the proposed

delineation between quantitatively—intense and qualitatively-diverse experiences in

16

organizations (Quinnones, Ford, & Teachout, 1995; Tesluk & Jacobs, 1998). This

approach is designed to overcome the narrow deﬁnitions from previous models of

experiential learning and focus on a more complete set of organizational experiences. As

outlined above, these distinctions build on recent advances in the micro-level study of

work experience and job performance, and have the potential to yield significant progress

in our understanding of the dimensionality of experience.

Table l — Different Types of Organizational Experience

 

 

Experience Type Deﬁnition Illustrations Examples
measurements of the .
Quantitative Intensity of time or amount of # of airplanes produced wr'gh" 1936

Experience

exposure to specific
situations or
environments
encountered over time

# of surgeries completed

Argote, et. al. 1990
Baum & Ingram, I998

 

 

Qualitative Diversity of
Experience

 

measurements of the
variety. challenge. or
complexity of the
situations or
environments
encountered over time

 

# of different past
alliance partners

Complexity of drugs
manufactured

 

Nerkar & Roberts. 2004
Hoang & Rothaermel. 2005
King & Tucci. 2002

 

First, it is important to be clear about definitions. To this end Table 1 is provided to

help clarify the nature of the proposed experience dimensions. According to Tesluk and

Jacobs (1998), experience can be described as consisting of quantitative and qualitative

components that exist at different levels of speciﬁcation and accrue over time.

Quantitative experience is defined as a measurement of the time or amount of exposure to

speciﬁc situations or environments encountered over time. Examples of quantitative

 

experience at the organizational level include number of years an organization has been
producing a certain product, number of years an organization has been part of a certain
alliance, the number of interactions the firm has had with a speciﬁc technology, the
number of patents it has applied for or any other additive representations of exposure
over time. These types of measures are reﬂective of the traditional measures of work
experience in the job performance literature (McDaniel, Schmidt, & Hunter, 1988), and
by translation, mirror the types of measures historically employed in the organizational
learning literature. In contrast, qualitative measures of experience include measurements
of the variety, challenge, or complexity of the situations or environments experienced by
individuals or organizations over time. Examples of qualitatively diverse organizational
experiences include the number of different partners that a ﬁrm has had alliances with.
the number of different types of technology a ﬁrm has worked with, the number of
different time-demand situations the ﬁrm has experienced, the difﬁculty of the ﬁrrn’s past
experiences, or the complexity involved in an organization’s past experiences. Important
to note is that an organization’s experience overcoming different types of challenges, in
different technological life-cycles, or with different consumer applications of
technologies are all qualitatively different experience sets that would not be identified
with traditional purely quantitative measures of organizational experience.

A key for the current research in examining these differences between various
types of experience is identifying how these experiences can lead to different learning
and ultimately to different innovation outcomes. In building their new model of the
different dimensions of experience, Tesluk and Jacobs (1998) also propose that the

different types of experience have different inﬂuence on learning from such experiences.

18

For example, individuals who perform the same task repetitively over long periods of
time gain extensive knowledge of the task and potentially develop short-cuts for task
completion (Tesluk & Jacobs, 1998). This allows the individual to become more efficient
over time and potentially more successful in their task environment. In contrast, Nemeth
( 1986) suggests that the exposure to qualitatively different technologies over time has the
effect of expanding an inventor’s cognitive models and can help them to make novel
connections that they may not have otherwise recognized. This is supported by ﬁndings
in the creativity literature that in many cases creative breakthroughs are associated with
the wide experiences of individuals with extensive social networks (Kelly & Caplan,
1997) who can draw on the ideas of diverse colleagues to make creative advances.

The important point here is that one type of experience (repetitive, quantitatively
extensive experience) can potentially lead to one learning outcome (efﬁciency) where a
different type of experience (diverse, qualitatively varied experience) can potentially lead
to a very different learning outcome (creative breakthroughs).

Extending Traditional Models of Experience Level

In addition to these different types of experience, major criticisms of research on
organizational learning through the years have focused on the confusion of multiple
levels of learning. In their inﬂuential work, Argyris and Schon (1978) suggest that the
dilemma of learning in organizations versus learning by organizations is the “paradox of
organizational learning.” They go on to argue:

It is clear that organizational learning is not the same thing as individual learning,

even when the individuals are members of the organization. There are too many

cases in which organizations know less than their members. There are even cases
in which the organization cannot seem to learn what every member knows (1978).

19

Further, this point has been echoed in every review of the organizational learning
literature since 1985 (Fiol and Lyles, 1985; Levitt & March, 1988; Huber, 1991;
Dodgson, 1993: Crossan, Lane, White, & Djurfeldt, 1995; Miller, 1996; Easterby-Smith,
1997; Bapuji & Crossan, 2004). Although exact agreements about the differences
between individual and organizational levels of learning are not well established (Cook
and Yannow 1993; Popper & Lipshitz, 1998), the one thing that is clear is that the
relations between individual and organizational levels of learning are complex. However,
there is growing evidence that these levels are both independently and jointly important
for ﬁrm-level learning and performance. For example, Hoang and Rothaermel (2005)
found that organizational-level learning from previous alliances inﬂuences the success of
future alliances. In this case, as might be expected, organizational-level experiences
inﬂuence organizational learning and organizational outcomes. On the other side, in a
recent theoretical piece Vera and Crossan (2004) propose a model where the individuals
in an organization (the CEO and top managers) play an important role in facilitating
learning at both individual and organizational levels of analysis. In this case it is the
individuals that are proposed to inﬂuence organizational-level learning and outcomes.
Thus, the notable conclusion from these findings is that organizational learning
involves a complex interaction of individual and organizational factors, and that new
models must address these factors. Further, of primary importance for future research is
the identiﬁcation of learning mechanisms that bridge the gaps between individual and
organizational levels (Cook & Yannow, 1993; Popper & Lipshitz, 1998) toward a more
complete model of organizational learning. Toward this end, the next stage of this

research will be to explore how the diverse types and levels of experience outlined here

inﬂuence the mechanisms for learning in the organization to guide organizational action,

innovation, and performance.

Table 2 — A New Framework of Organizational Experience

Nature of the Experience

 

Quantitative Intensity

Qualitative Diversity

 

Individual

Individual Quantitatively
Intensive Experience

e. g. # of years of experience in
semiconductor R&D jobs

Individual Qualitatively
Diverse Experience

e. g. # of different semiconductor
technologies exposed to in past jobs

 

Organizational

Level of the Experience

 

 

Organizational Quantitatively
Intensive Experience

e.g. # of years designing and
producing semiconductors

 

Organizational Qualitatively
Diverse Experience

e. g. # of different electronics
technologies used in R&D divisions

 

A New Multi-Level Model of Organizational Experience

Bringing the ideas from the previous two sections together, I build on these

foundations to propose a new integrative model of organizational and individual

experience. Although the differences between diverse types of experience at multiple

levels combine to potentially highlight many different paths for research, the core of

these distinctions can be integrated into a 2 x 2 matrix of key experience constructs for

the learning and innovation of organizations. A graphical representation of this new

model is provided in Table 2. By synthesizing the ideas from these many different

experience perspectives the result is a more complex view of experience and learning

21

 

than that of the simplistic quantitative models of traditional literature', and a more
nuanced set of constructs for exploring firm processes and outcomes. Further, this new
view of experience helps to inform the next section of this research, an exploration of the

organizational learning mechanisms inﬂuenced by these different experiences.

APPLYING THIS NEW MODEL OF EXPERIENCE TO SEARCH AND LEARNING

The Mechanisms of Organizational Learning

Thus, with this new theoretical model of organizational experience, the next step
for this research is to apply this new experiential model to how organizations actually
learn. To do this I will investigate the mechanisms for learning through which these
experiences are embedded into the organization, repeatedly transformed into
organizational action, and potentially result in radical innovation. How do organizations
integrate new knowledge from experiences? Although models describing the mechanisms
of organizational learning have been slow to develop in the literature, several theoretical
advances inform this process.

Notably among these is Kolb’s 1984 book on experiential learning by individuals.
In this research Kolb draws on a long history of research in psychology to present a well-
researched model through which experiences are translated into action. Although debate
continues as to whether or not to directly adopt this model for organizational-level

experiential learning (Popper & Lipshitz, 1998), at the very least it provides important

 

' One point to note here is that technically. both qualitative diversity and quantitative intensity assessments
of experience are likely to be quantitatively assessed in measures of these different types of experience. In
this sense. the use of the term ‘qualitative' in describing the diversity of experiences does not mean that the
assessments of such diversity would necessarily be assessed qualitatively. just that the measurement of
experiences could assess the amount of quantitative intensity (repeated similar encounters) in experience or
assess the amount of qualitative diversity (exposure to different types of encounters) reflected in previous
experiences. even if both measures are actually measured quantitatively.

22

insights into the process. In his model, Kolb (1984) suggested that individuals experience
events, reﬂect and make generalizations about these experiences, and then experiment
and make changes to the structure and nature of their behavior in light of the new
knowledge they have acquired. Once these insights from experience have been embedded
into the individual’s routines and behaviors then this provides a new foundation from
which new experiences can again change and re-shape the framework (Kolb, 1984). This
general process is echoed in recent theoretical advances from Crossan, Lane, and White
(1999) where organizational learning is conceptualized as involving stages of intuiting
and interpreting by individuals, integrating by groups, and institutionalizing by
organizations. In both of these models we see evidence of a general process that involves
experience, reﬂection, and then change. In particular, two key contributions from
Crossan, et. al. (1999) add specific insight into the process. First, organizational
processes occur at multiple levels and each of these levels contributes to changes in the
structure and processes of the organization. Second is the idea that organizations also
have an institutional level of norms, a level not as clearly present in individuals. This
institutional level is particularly important as it is both inﬂuenced by experiences and
endures over time to act as a lens through which future experiences are evaluated by the
organization (Crossan, et. al., 1999; Daft & Weick, 1984). In a sense, this institutional
level informs not just what individuals in organizations do but also why they do it.
Further supporting this idea that experiences are preserved and inﬂuence future
interpretation are notable ﬁndings from several authors. In this stream it is argued that
organizations, unlike individuals, develop and maintain learning systems that not only

inﬂuence their immediate members, but are then transmitted to others over time by way

of organizational histories and norms (Lawrence & Dyer, 1983; Levitt & March, 1988;
Popper and Lipshitz, 1998). Hedberg (1991) notes, “members come and go, and
leadership changes, but organizational memories preserve certain behaviors, mental
maps, norms, and values over time.” The most prominent contribution to these ideas of
organizational memory comes from Walsh and colleagues. Walsh (1995) argues that in
addition to guiding behavior and organizational action, learning mechanisms act as
memory devices, through which the accumulated knowledge of the organization is used
to repeatedly enact processes over time. Walsh and Ungson (1991) argue that experiences
are interpreted and then stored throughout the organization, in individuals, culture,
transmissions, structure, and ecology (physical layout). In line with these ideas, Levitt
and March (1988) suggest that organizations can be seen as learning by encoding
inferences from history into two primary mechanisms, routines that guide typical
behavior and norms that govern “the way things are done around here.” However, these
authors also note that relatively little is known about the details through which experience
is accumulated into such structures (Levitt & March, 1988).

Building on and integrating these ideas, the current research focuses on the
inﬂuence of experiences on the selection, usage, and adaptation of norms and routines in
the organization. Building on the work of Nelson and Winter (1982) and others, routines
are deﬁned as repetitive and recognizable patterns of interdependent actions that govern
the typical enactment of a speciﬁc task or set of tasks in an organization over time
(Feldman & Pentland, 2003; Pentland & Rueter, 1995). In contrast to routines,
organizational norms are defined at a higher, more institutional level as the common

beliefs about appropriate and required behavior for organizational members as members

of the organization (Katz & Kahn, 1978), a definition that builds on earlier foundations
from Sherif (1936), Davis (1950) and others. Where routines directly inﬂuence
organizational members by directing their activity in the context of task completion,
organizational norms indirectly impact individuals by biasing them toward certain types
of organizationally-accepted decisions and ways of working (Selznick, 1957).

In this sense, experiences are viewed as inﬂuencing both the organization as a
whole and individuals within the organization, and the learning that occurs as a result of
these experiences is encoded into the norms and routines of the organization. These
norms and routines then store and preserve knowledge over time. As the ultimate
outcome of interest in the current research is new knowledge creation, this focus on
norms and routines helps to answer the question of how some organizations can sustain
success over long periods of time. As knowledge creation occurs within a normative
context (Knorr—Cetina, 1981; Dougherty & Heller, 1994), replicating the success of past
innovation requires learning and repeating, which is facilitated by establishing norms and
routines. To this end, organizational norms and routines are two facets of the organization
that endure and continue to guide the organization through product cycles, CEO’s, and
technological advances. Although star performers and great inventors can fade away, lose
their edge, or be hired away by competitors, organizational norms and routines carry on
through generations of employees (Harrison & Carroll, 2006).

The Critical Role of Search Routines in the Pursuit of Radical Innovation

Building on these ideas, a key question remaining is that of which norms and

which routines contribute to the development of radical innovations. Although the norm

and routine literatures suggest links between experience and these processes. to

25

understand how these processes inﬂuence radical innovation we must identify the
speciﬁc norms and specific routines that inﬂuence radical innovation. To this end, several
recent advances provide guidance.

Specifically, a growing literature has built up over the last ten years to suggest
that the routines organizations use to search for information in their research and
development projects greatly inﬂuences the nature of the innovation (Stuart & Podolny,
I996; Ahuja, & Lampert, 2001; Fleming, 2001; Rosenkopf & Almeida, 2003). This
search research builds on the theoretical tradition of the Carnegie school and the
behavioral theory of the firm, which suggests that search processes are important
determinants of organizational outcomes, particularly innovation (Cyert & March, 1963;
March & Simon, 1958). Further, there is considerable agreement in the literature that at
the organizational level, local search is often produced through the smooth functioning of
organizational routines (Stuart & Podolny, 1996; Nelson & Winter, 1982; Cyert &
March, 1963). Recent empirical and simulation research has continued this tradition by
examining how organizations build on their own experience sets (Baum, Li, & Usher,
2000) to search for information and adapt to rugged landscapes in efforts to improve
ﬁtness (Gavetti & Levinthal, 2000; Levinthal, 1997). These search routines can involve
investigating narrow or broad sources of information (Rosenkopf & Almeida, 2003),
information primarily from a ﬁrm’s own previous innovations (Katila & Ahuja, 2002),
old or new knowledge (Nerkar, 2003; Katila, 2002), or information only from proximal
and familiar sources (Ahuja & Lampert, 2001). In this line, of critical interest is how
organizations can overcome local search biases through the use of different search

routines (Rosenkopf & Nerkar, 2001) or knowledge strategies (Bierly & Chakrabarti,

26

1996) to achieve radical innovation. However, despite these significant advances in the
modeling and understanding of search routines, little is understood about how different
types and levels of organizational experiences inﬂuence the selection, usage, or
development of search routines. There is evidence that experience inﬂuences search
(Ahuja & Lampert, 2001: Nerkar & Roberts, 2004), but which experiences and how
remain unclear.
Supporting and Conﬂicting Norms

Although such search routines are argued to have direct inﬂuences on the creation
of radical innovations in organizations, like most institutional pressures, the inﬂuence of
organizational norms is likely to be more indirect. Rather than directly contributing to the
creative effort of producing innovation, norms support or constrain which experiences
occur and are drawn upon to select and develop routines. Research on institutional theory
suggests that individuals interpret their experiences through the ﬁlter of the institutional
environment in which they work (Scott, 1995). This implies that although individual
experience may dictate one type of information search behavior, the institutional
pressures they face may advise an alternative choice. For example, so-called “cultures of
efﬁciency” such as those advocated by TQM and six-sigma programs (Westphal, Gulati,
& Shortell, 1997) support economic goals in organizations by focusing organizational
members attention on the fast and efﬁcient completion of tasks (Camerer & Vepsalainen,
1988). In this type of an efﬁciency-driven normative environment. the conversion of
individual experiences into action is likely to be biased toward more efﬁcient search
routines. If individuals who enter this environment have wide and diverse experiences

that lead them to engage in extensive search and consider a broad set of information

27

sources in their innovative efforts, a conﬂict emerges. On the other hand, other
organizational cultures are focused on supporting creative thinking and potentially non-
task-related social networking to stimulate novel ideas. Examples of such creativity-
focused organizations include the well-documented imaginative environments of Apple
Computers, Pixar Animation Studios, IKEA, or MTV Networks (McGregor, 2008). In
these creativity-based normative environments the conversion of individual experiences
into action will be biased toward more creative and diverse information search routines.
Individuals who have historical experiences more focused on efﬁciency or narrow
domains will again ﬁnd themselves in a search conﬂict.

Thus a variety of questions emerge. Organizations have experiences that, over
time, directly contribute to the development of institutional norms and organizational
structures to support these norms (Meyer, Scott, & Deal, 1981; Scott, 1995). Individuals
inside organizations have experiences that are both unique from and shared with those of
the organization, and may have experiences that lead them to engage in routines that are
out of alignment with the pressures of the normative environment. Given the importance
of search for the creation of radical innovation (Ahuja & Lampert, 2001), how these
divergent experience sets inﬂuence the selection, usage, and development of search
routines is likely to significantly inﬂuence the long-term success of organizational
innovation efforts.

THEORY AND PROPOSITIONS
A New Model of Experience. Search. and Knowledge Creation
Figure 1 presents the full proposed conceptual model of experience, search, and

new knowledge creation. In building propositions to support this model, I will begin by

examining how different types of individual level experience inﬂuence the choice of
search routines. From this foundation, I will move on to examine how the different search
routines modeled relate to each other. Next I will bring in the organizational level of
experience, where different types of organizational experience are posited to inﬂuence
both the norms and routines of the organization. Building on these ideas I move to the
moderating inﬂuence of organizational norms on the relations between individual
experience and search routines. Finally I examine the relations between organizational

search routines and innovation.

29

Figure 1 — Full Conceptual Model of Experience, Search, and New Knowledge Creation

. _ a...» El _ «0550m— :0..an

_ a\u:l

  

   

Co 5805

p ......................... .

. 3 >o:o_o_tm _
P .........................

      

_ vocatoaxm
_ 2.3.30 9o

p ....................... _
H 3 on

EoEco.__>:m 02.2502
.0 cozaucﬂuoﬁﬁﬁuo;

  

 

 

 

 

m .
.............. . .w L

at ”I ....................... I”

1 35:33 _

3 «a _ «25:20 in... m

P .........................

< _ macaroaxm m

, _o>.a_ca: .>.cu

N1 «+» kl .p ..... . m..u.lu::lm~....:.lm_l_

 

m n
m EoEco=>cm 035.502 _

m .3 :o_.ﬁco_..o.>~_>=o:uo.i _

HER
0023.593 _
_ 2:22:30 90 m
—. .........................

30

Individual Experience and Organizational Search Routines

Building on the review of the search literature, the first relations in the proposed
model explores the link between individual experience and the search routines used by
individuals in their research investigations. Extending the research on individual
quantitative experience from the micro-level tradition, higher levels of individual
quantitative experience are posited to lead to a tendency toward the selection, usage, and
development of more efficient search routines. Efficient search routines are deﬁned as
processes that exist in organizations to guide typical action toward exploration for
relevant information that attempts to be as fast and competent as possible given
constraints. These experiences increase the tendency toward efﬁcient search routines in
three ways. First, research on individual career experience has shown that with increasing
levels of quantitatively-intensive experience individuals develop enhanced job knowledge
(Schmidt, Hunter, & Outerbridge, 1986), which allows the individual to develop in-depth
tacit knowledge about the best ways to do their work (Berman, Down, & Hill, 2002).
This extensive job knowledge helps individuals to develop short-cuts and heuristics
which make future search investigations easier, faster, and more efﬁcient. Second,
increasing levels of individual quantitative experience allows individuals to develop
extensive knowledge of where to ﬁnd important new information. In research on teams
this type of knowledge is known as transactive memory, or the knowledge of where
knowledge resides in a team (Wegner, 1986). In searching for other information the same
type of where-to-find-it knowledge is developed, such that with extensive experience
individuals learn where to locate information they need to solve problems that are similar

to problems they have solved in the past. This transactive memory has been shown to

31

improve performance and efficiency (Austin, 2003), saving individuals from the need to
try out different tools in an exploratory model of trial-and-error. Finally, increasing levels
of individual quantitative experience also allow individuals to become increasingly adept
in their search. This advanced user knowledge allows individuals to avoid the pitfalls and
frustrations of first-time search (Compeau, Higgins, & Huff, 1999; Burkhardt & Brass,
1990) and proceed in search activities efﬁciently with confidence that achieved search
results reﬂect the best use of tools available. Based on these ideas, I propose:

Proposition 1: Increasing levels of quantitatively intense individual experience

will lead to the selection, use, and development of more eﬁicient organizational

search routines.

Examining an alternative perspective of individual experience (from Table 2),
qualitatively diverse individual experiences are posited to have a different effect on
search. In contrast to the efficiency beneﬁts of narrow experience, the qualitative
diversity of experiences is proposed to be associated with the selection and development
of broader search routines. Broad search routines are deﬁned as processes that exist in
organizations to guide typical action toward exploration for relevant information that
attempts to be as comprehensive and inclusive as possible given constraints. This broader
effect on search is posited for several reasons. First, diverse experiences expose the
individual to diverse sources of knowledge and a wider range of potential tools to use in
search. This knowledge of diverse sources of information provides such individuals with
a wider ‘evoked set’ (March & Simon, 1958) of available alternatives when faced with
new problems. As such individuals are able to draw on a wider variety of search aids to
investigate a new research question, and are likely to consider more potential sources of

knowledge than those with more limited experience. Second, a diverse set of individual

32

experiences exposes the individual to alternative cognitive models (Pitcher & Smith,
2001). These unique cognitive models provide a variety of distinct lenses through which
to consider new problems, and may lead to alternative framings of new problems. As
these more novel cognitive framings are explored, a wider set of information is necessary
to pursue the ideas, leading to wider search. Finally, a diverse set of experiences also
helps the individual to build more advanced individual skills in the use of a variety of
different search methods. A major obstacle that hinders the breadth of search is the
knowledge of where to ﬁnd new information and how to use the investigative tools
needed for such wide search (Rosenkopf & Almeida, 2003). Because individuals with
advanced skills in the use of such tools do not face such obstacles, they should be more
willing and able to search broadly. These ideas come together to support the following
proposition:

Proposition 2: Increasing levels of qualitative diversity in an individual ’s

experiences will lead to the selection, use. and development of more breadth-
oriented organizational search routines.

Organizational Experience and Organizational Search Routines

Moving now to the other side of experience in organizations, the next set of
relations in the proposed model explores the link between organizational-level experience
and search. In this part of the model, increasing levels of quantitatively-intensive
experience encountered by the organization are again argued to be associated with a
focus on the selection, usage, and development Of more efficient search routines in the
organization. Several findings support this conclusion. First, similar to the dynamics at

the individual-level (Schmidt, et. al, 1986), increasing levels of quantitative experience at

33

the organizational-level allow for the development of tacit knowledge, search short-cuts,
and helpful search heuristics. However, because these efﬁciency improvements and
short-cuts are learned at the organizational level, they can be institutionalized into
standardized procedures (Pugh, Hickson, Hinings, & Turner, 1968). Where individual
quantitative experience provides employees with the opportunity to develop such helpful
tools, organization-level experience provides the opportunity for such short-cuts and
heuristics to be standardized and made widely available, known, and accessible, helping
to increase the efﬁciency of future search. Second, increasing levels of organizational
experience facilitate the development of centralized information and search centers,
knowledge management systems, and knowledge repositories (Davenport & Prusak,
1998). In this sense, an organization that has a long set of experiences and a tradition of
engaging in specific types of search activities is likely to develop centralization systems
to make knowledge widely available, providing a convenient and proximal search
environment for future searches. This helps to increase the efﬁciency of such searches.
Finally, increasing levels of organizational experience also facilitate the development of
organizational-wide competences (Orlikowski, 2002; Henderson & Cockbum, 1993;
Snow & Hrebiniak, 1980), in this case in the knowledge and skilled use of certain search
tools. These skills are then perpetuated over time, as organizations with such widely
distributed skills in search become very good at training new employees and developing
employees in the efﬁcient use of these tools (Feldman, 1981; Chao, O’Leary-Kelly, Wolf,
Klein, & Gardner, 1994). Building on these ideas, I propose:

Proposition 3: Increasing levels of quantitatively intense organizational

experience will lead to the selection. usage. and development of more eﬁ‘icient
organizational search routines.

34

Moving again to a qualitative view of experience from the new experience matrix
(Table 2), the variety, challenge, and complexity of an organization's experiences are all
posited to have the opposite effect on organizational search. Specifically, diverse
organizational experiences are argued to positively inﬂuence organizational search
breadth in at least three ways. First, increasing levels of diversity in an organization’s
historical experiences allow the organization to develop diverse sources of knowledge
inside the organization (Grant, 1996). In this sense, over the years the wide variety of
experiences will have forced the organization to develop systems and tools for
investigating the many different technologies or services in their product portfolio. Over
time this facilitates the development and institutionalization of a wide variety of
information search tools. These diverse knowledge tools can then help to provide access
to a wide range of knowledge domains and resources internal to the ﬁrm to assist future
search activities. Second, organizations with a wide set of experiences have the
opportunity to develop multiple cognitive models among individuals inside the
organization. Cognitive models, or ‘dominant logics’ (Prahalad & Bettis, 1985), can be
created over time in response to the tradition of collective experiences in the ﬁrm. When
these experiences are narrow, restrictive and focused cognitive models can lead to the
development of problematic core rigidities (Leonard-Barton, 1992). However, when
experiences are wide, diverse, and distributed throughout the organization, such
restrictive models can be avoided and instead intellectual breadth and respect for a
diversity of ideas can be fostered. This makes a wide variety of diverse perspectives,
cognitive models, and alternative solution sets available to future search investigations -—

helping to foster broad search routines. Finally, because the organization has a wide base

35

Of experience, this allows individuals in the organization to develop a strong collective
competence (Orlikowski, 2002) in use of multiple search tools and methods in-house.
Research has shown people often suffer from ‘spatial myopia,’ (Levinthal & March,
1993), tending to rely on information in their proximal environment rather than expand
out to more diverse potential sources. In fact, in a study of building layouts, Allen (1977)
noted that researchers in laboratories often restrict their search to individuals located near
them in geographic space and that increases in distance from just 10 to 50 meters resulted
in a serious drop in communication. Instead, people focus on the resources available
around them. Because organizations with wide experience sets have diverse skills and
knowledge sets all located in a socially proximal environment (in-house), this makes it
easier and more likely for individuals to engage in wide future search. Based on these
arguments, I propose:

Proposition 4: Increasing levels of qualitative diversity in an organization ’3

experiences will lead to the selection, use, and development of more breadth-

oriented organizational search routines.
Organizational Experience and Organizational Norms

Moving now to the inﬂuence of organizational-level experiences on
organizational norms, several relationships are proposed. To begin with, increasing levels
of quantitatively-intensive experiences are argued to lead to the development of
institutionalized norms of efficiency in at least three ways. First, with more
quantitatively-intensive experience an organization’s entire focus can be directed at the
perfection of a certain set of tasks and technologies. An organization that is new to a
technical area needs to experiment to try and ﬁnd the best ways to do things. In contrast,

an organization with extensive experience can simply focus on optimal execution. This

36

narrow focus helps to increase efficiency by reducing distractions, helping to create an
organizational competency for efficient task execution. As corporate cultures and
normative structures tend to be built around what an organization does best, and does
most often (Feldman, 1984: Martin, 2002), norms are likely to be created to support this
competency.

Second, organizations with extensive quantitative experience often develop
implicit cultural systems to support efﬁciency. One such cultural system is the
development in many organizations of informal practices that support the dominant
culture (Martin, 2002). For example, at General Motors in the 1960’s a standardized
efﬁciency mentality was evident in everything from the assembly lines to the corporate
executive suites, where standardized dress codes and office décor were restricted to
muted tones and minimalist designs (Martin & Siehl, 1983). Such informal systems help
to reinforce expectations and keep the normative focus on efﬁciency throughout the
organization. Further, these practices make an efficiency expectation and norm more
ubiquitous and institutionalized. Finally, over time experience and efficiency orientations
tend to become supported by the development of supervisory systems focused on
efficiency (Patchen, 1962) and compensation systems that reward efﬁcient outcomes
(Jansen & Von Glinow, 1985). As employees are motivated to do what they are told and
paid to do (Yukl, 1998; Vroom, 1960), such systems help to create a normative
environment supporting efﬁcient execution. Based on these ideas, I propose:

Proposition 5: Increasing levels of quantitatively intense organizational

experience will lead to the development of more productivity-oriented
organizational norms to support the eﬂicient execution of organizational tasks.

37

Next, variety in an organization’s experience set is argued to lead to very different
organizational norms. In this case, several ideas and findings support the argument that
diversity in experience at the organizational-level will lead to the development of more
creativity and breadth norms in the organization. First, with a greater diversity of
experiences, the organization can foster the development of diversity in its communities
of learning (Brown & Duguid, 1991) such that many different ideas and cognitive
frameworks are respected (Fiol, 1994). In narrow, single-business organizations it is easy
for a narrow set of execution norms to take center stage. However, in more broad and
widely experienced organizations no single ideological orientation has the ability to
hijack the corporate norms (Raelin, 1997). Instead this diversity provides space for a
more interactive and creative discourse to become the pervasive mentality for knowledge
creation endeavors (Nonaka, 1994). In organizations where no single set of ideas reign
supreme, breadth becomes part of the culture and diversity becomes the norm (F iol,
1994). Further, research has suggested that information technology may be created in
such organizations to support diverse community-level hermeneutics and interpretive
processes (Boland, Tenkasi, & Te’eni, 1994). As corporate cultures and normative
structures tend to be built around the consensus values of the organization (Feldman,
1984; Martin, 2002), the lack of a dominant coalition will lead to norms that support the
openness of the community to new ideas.

Second, organizations with diversity in experience are also likely to develop
implicit cultural systems to support the openness of their culture. As in the previous
section, an important part of such cultural systems are the development of informal

practices that support the culture (Martin, 2002). For example, at Hallmark (the greeting

38

card company) the artistic creativity mentality is so predominant that employees are
encouraged to find a quiet place for reﬂection outside of work in a park or nature center
that they can go to occasionally for inspiration when they are having trouble thinking of
new ideas (Shields, 2005). This type of informal system helps to reinforce openness and
creativity expectations in the workplace and keep the normative focus on broad
imagination throughout the organization. Further, these practices make breadth as an
expectation and norm more ubiquitous and institutionalized. Finally, over time these open
orientations tend to become supported by the development of supervisory systems
(Patchen, 1962) that encourage breadth and compensation systems that reward integrative
outcomes (Jansen & Von Glinow, 1985). As employees are motivated to do what they are
told to do and paid to do (Yukl, 1998; Vroom, 1960), such systems help to create a
normative environment supporting broad thinking and integrative pursuits. Based on
these ideas, I propose:

Proposition 6: Increasing levels of qualitative diversity in an organization’s

experiences will lead to the development of more learning-oriented

organizational norms to support the creative pursuit of organizational initiatives.
The Moderating Inﬂuence of Organizational Norms

Moving on to the inﬂuence of organizational norms, two relationships are
proposed. To begin with, organizational productivity norms are argued to moderate the
effects of individual experience on routine selection, use, and development. In this sense,
although Proposition 2 suggests that the relation between individual experiences and
search is such that more qualitative experience is expected to lead to breadth in search
routines, in the presence of strong productivity norms more efﬁcient search routines are

argued to become adopted. Several ideas and research findings support this proposition.

39

Most notably, both laboratory and field research provide evidence that in situations where
the individual experiences and biases are out of ﬁt with the accepted culture, such
individuals either adapt to the culture (Francis, 1991), suffer from poor performance
(O’Reilly, Chapman, & Caldwell, 1991), or leave the organization (Vandenberghe,
1999). Providing support for this idea, McNamara and Bromiley (1997) found in a study
of 223 commercial lenders that when making risk assessments of customers,
organizational factors overwhelmed the inﬂuences of individual cognitive biases. In the
case of individual biases toward technological search, although individuals may desire to
engage in wide search it is likely that they will have difficulty overcoming social pressure
(J aeger, 1983) from peers and managers to be more productivity-oriented. Further adding
to this mis-alignment problem is the fact that in a strong productivity—oriented culture
jobs are often designed speciﬁcally to support and require efﬁciency, in an effort to
increase task focus and conformance to organizational standards (Camerer &
Vepsalainen, 1988). Based on these restrictive job designs, individuals who have an
individual bias toward increased breadth in their information search would need to
engage in extra role behaviors to find search tools beyond the few tools available. These
factors combine to create a hostile and difﬁcult environment in which to engage in broad
search and lead to the overall institutional pressure for conformity. On the other hand, in
the case of individual and organizational experience ﬁt, such normative support should
help to strengthen the relationship between quantitatively intensive experience and
efﬁcient search by providing a strong and supportive environment for efficient action.

This leads to the following proposition:

40

Proposition 7: Productivity-orientation in organizational norms will moderate the
relation between individual experience and search such that these normative
pressures will bias individuals toward less broad and more eﬂicient search.

In contrast to the normative environment of productivity, in other organizations
norms encouraging breadth and integration are argued to overwhelm the effects of
individual experience on routine selection, use, and development. In this sense, where in
Proposition 1 the relation between individual experiences and search is such that more
quantitative experience is expected to lead to efﬁciency in search routines, in the
presence of breadth norms broader search routines are argued to become adopted. Again,
several ideas and research ﬁndings support this proposition. As outlined above,
signiﬁcant evidence suggests that in situations where the individual experiences and
biases are out of fit with the accepted culture, individuals tend to either adapt to the
culture (Francis, 1991), suffer from poor performance (O’Reilly, Chapman, & Caldwell,
1991), or leave the organization (Vandenberghe, 1999). In the case of individual
tendencies toward search, although individuals may desire to engage in the narrow and
efﬁcient search that they are accustomed to, they will have difﬁculty overcoming social
pressure (J aeger, 1983) from peers and managers to search more broadly. Further adding
to this argument is the fact that in a creativity-oriented culture jobs are often designed
speciﬁcally to support and require extended search in an effort to increase breadth and a
culture of integrative thinking (Amabile & Conti, 1999; Blau & McKinley, 1979). Based
on these integration-minded job designs, individuals who have an individual bias toward
narrow and efﬁcient focus in their information search would have to essentially ignore
the diverse set of tools made available to and expected to be used by each employee. For

example, one can imagine a R&D team progress report meeting where each team member

41

presents the details and results of their broad information search to the team. Individuals
who present search results without an explanation of rejected alternatives from broad
search would likely face a hostile and questioning set of peers whose expectations for
breadth would have been unfulﬁlled. These factors combine to create a difficult
environment in which to engage in only narrow search and lead to institutional pressures
for conformity. On the other hand, in the case of individual and organizational experience
ﬁt, such normative support should again help to strengthen the relations between
qualitatively diverse experience and broad search by providing a strong and supportive
environment for creative action. This leads to the following proposition:

Proposition 8: Breadth-orientation in organizational norms will moderate the

relation between individual experience and search such that these normative

pressures will bias individuals toward broader and less efficient search.
Connecting Search Routines to Innovation

Finally, in the last section Of the model the different types of search routines
developed are posited to lead to very different innovation outcomes. Starting with
efficiency, several findings support the proposition that efﬁcient search routines lead to
an increased rate of incremental innovation in organizations, but not breakthroughs. First,
the efﬁcient use of human capital in this search paradigm allows the same research
personnel to do more with their time and produce more innovations. However, because
this search is efficient in large part due to the lack of integration of diverse sources of
information, such search is unlikely to bring together novel solutions in the ways
necessary for the achievement of radical innovation (Ahuja & Lampert, 2001). Second,
the rate of innovation is enhanced by the faster tumaround on bad ideas that is possible

when search does not need to be as wide. This makes it possible for design problems to

42

be solved faster (Fleming & Sorensen, 2004). This increased speed in problem solving
will help support a more rapid schedule of producing innovations at a more robust pace
than operations that focus on wide search. Third, research has clearly demonstrated that
certain structures support incremental innovation, where others support radical innovation
(Downs & Mohr, 1976). For example, Ettlie, Bridges, and O’Keefe note that in a study of
firms in meat, canning, and fish industries that structures used to support efficiency (i.e.
formalization) were positively correlated with the rate of innovation in such ﬁrms, and
more strongly associated with incremental innovation than radical innovation (1984:
692). Finally, because efficient search focuses on narrow and familiar sources of
information, the consistency of such knowledge with the investigator’s existing
knowledge paradigms helps to make such knowledge more easily accessible.
Speciﬁcally, this familiarity allows researchers to make more efficient use of existing
knowledge in the pursuit of incremental innovation (Fleming & Sorensen, 2004). These
arguments lead to the following proposition:

Proposition 9: Increasing levels of eﬂiciency in organizational search routines

will be associated with an increasing rate of innovation by the organization.

Next I explore the culmination and ultimate goal of the proposed model in the
creation of radical innovations in organizations. To this end, breadth in organizational
search routines is argued to support radical innovation for several reasons. First, in
studies of creativity and the creation of breakthroughs, creativity has generally been
found to result from the synthesis of disparate and novel ideas (Simonton, 2002). For
example, in a narrative on the development of Hewlett-Packard’s thermal ink- jet

technology, Fleming (2002) found that it was the integration of diverse ideas from

43

electronics, ﬂuid dynamics, and physics that led to the successful creation of this
breakthrough. Second, this idea is supported by research which has suggested that it is in
the exposure to a wide and diverse set of ideas that allows researchers to form new and
innovative cognitive models which help to conceptualize problems differently (Nemeth,
1986). This broad exploration and unique perspective then acts as a foundation for radical
innovation (Rosenkopf & Nerkar, 2001; Fleming, 2001). Without the out-of—the-box
thinking that is induced by wide search, discovery of path-breaking new products and
services is an unmanageable process of chance and good fortune. However, with the
management of search routines to make the integration of novel ideas common and the
bending of cognitive models routine, radical innovation becomes the natural and expected
outcome of sound managerial practice (Fleming, 2002). Based on these arguments, I
propose:

Proposition 10: Increasing levels of breadth in organizational search routines

will be associated with an increasing rate of radical innovation by the

organization.

Finally, the last propositions argue that these relations not only act to support very
different innovation outcomes, but that they also actively act to suppress each other. In
this case increasing efﬁciency in organizational search routines is proposed to be
associated with fewer breakthrough innovations and increasing breadth in organizational
search routines is proposed to be associated with a slower rate of overall innovation.
These cross-negative relations are proposed for several reasons. First, in breadth—oriented
organizational search routines the focus is on exploring a wide and varied set of
information sources to ﬁnd an ideal solution. Increasing the rate of innovation requires

that search be completed quickly, which sparks a conﬂict. The time spent searching

through a diverse set of alternatives will act to directly impede the speed and efficiency
of innovation efforts, leading to a slower rate of new innovation. In addition, knowledge
of a diverse set of information sources may make it difﬁcult for researchers to focus
narrowly on a single set of ideas or narrow projects. Supporting this idea, research on
information processing has shown that in teams where efﬁciency is expected, providing
access to diverse information sources impedes the efﬁciency of team performance due to
information overload (Katz & Tushman, 1979; Keller & Staelin, 1987).

Proposition Ila: Increasing levels of breadth in organizational search routines

will be associated with a decreasing rate of innovation by the organization.

However, on the other side, routines which are focused on efficiency are likely to
work against the conditions that support creative breakthroughs. Although creative
breakthroughs are impossible to speciﬁcally plan, conditions that support such radical
advances include the integration of novel and unconnected ideas (Ahuja & Lampert,
2001) as well as the exposure of researchers to alternative cognitive models (Nemeth,
1986; Pitcher & Smith, 2001). Scientists who work quickly and efﬁciently by focusing on
narrower search routines would be less likely to be exposed to these new ideas or
alternative cognitive structures and thus less likely to find or pursue radical or novel new
ideas. Together, these arguments support the following proposition:

Proposition 1 I b: Increasing levels of efﬁciency in organizational search routines

will be associated with a decreasing number of radical innovations by the

organization.

DISCUSSION AND CONCLUSIONS
The consistent theme throughout this endeavor has been the idea that key

differences exist between quantitatively-intensive and qualitatively—diverse experiences,

45

and further, that such differences have significant influences on organizational norms and
routines. In moving forward from more traditional and narrow definitions of experiential
learning, this research has the potential to open new doors both for future theoretical and
empirical research.
Implications of this Research for Organizational Learning Theory and Practice

To begin with, this research has implications for theory in the domain of
organizational learning. First, this research highlights the fact that we are likely reaching
the end of traditional research on learning curves and the benefits to firms that accrue to
increasing levels of similar experience. However, this paper does not close the door on all
experience-based organizational learning, it simply re-directs such research toward a
focus on qualitative diversity in organizational experience. In fact, I would argue that the
most fruitful directions for continued research in organizational learning lie in further
breaking apart qualitative diversity in experience and the very different effects such
diversity has on organizational choices and performance. This research provides both the
direction and some tools for moving in these new directions. Second, this research takes
initial steps into better considering the antecedents of search in organizations. After Cyert
and March (1963) initially introduced the idea of organizational search into the
management literature, search research moved forward in small steps until recently.
However, in the last 10 or 15 years, such research has greatly expanded (Ahuja, &
Lampert, 2001; Katila, 2002; Katila & Ahuja, 2002; Nerkar, 2003; Rosenkopf &
Almeida, 2003; Stuart & Podolny, I996). The current research provides an important
extension to this new stream of search research by developing theoretical arguments for

the experiential roots of such search to demonstrate how such search may be shaped and

46

at times path-dependently constrained in many organizations. Third, this research has
implications for research on norms and routines. In fact, despite years of research on
norms and routines in the organizational literature (Nelson & Winter, 1983; Stuart &
Podolny, 1996; Levitt & March, 1988; Feldman, 1984), we are only beginning to learn
more about how such core processes are formed and re-created over time in
organizations. This broader conceptualization of organizational experience provides a
first step toward better understanding how these fundamental organizational processes are
created and shaped over time. This challenges simple definitions of the experiential roots
of norms and routines (Bettenhausen & Mumighan, 1985; 1990; Stuart & Podolny, 1996)
to suggest key differences in the effects of different types of experience on norm and
routine development.

In addition to implications for theory, this research also has several potential
implications for practice. First, this research provides a model for helping managers
understand the path dependencies that explain why their business may be stuck in a
certain type of knowledge creation mode. Although the development of rigidities may be
difficult to trace over time, the model proposed in this research provides a method for
managers to begin making attributions to understand accrued path dependencies. Second.
this research provides guidance for how organizations can take steps to shape their norms
and routines. Specifically, by pushing into the sort of black box of antecedents to norms
and routines, this research breaks apart the experiential roots of such processes to suggest
how future change management programs can focus on the development of different
types of experience to support different norms and routines. Finally, this research also

provides a new way for managers to build up organizational learning initiatives toward

47

 

the types of innovation they would prioritize for their organization. Although many
executives preach learning to their stakeholders, the ideas in this research provide a path
to aid the development of specific types of learning competencies over time.

Future Directions

In terms of future directions, this research also highlights several potential paths.
First, additional theory is needed on the differential inﬂuences of qualitatively diverse
experience. In exploring the antecedents of different types of search routines in this
research, I simply argued that any type of qualitatively diverse experience would have the
effect of broadening individual and organizational approaches to search. Although this
broad classiﬁcation made sense for this research, where exposure to new ideas and
methods was key, in other contexts splitting the effects of experience variety, challenge,
and complexity could have interesting consequences for theory development. For
example, individual and organizational experiences that are challenging could have
different effects, both at different levels of analysis, and on potential outcomes.
Specifically, depending on the resiliency and internal fortitude of such individuals and
organizations to persist after failure, challenge may lead to either increased or decreased
motivation. These represent potentially fruitful theoretical extensions of the proposed
model.

In addition, this research also points to the need for future empirical research on
experiential learning. Although direct empirical testing of the proposed model (Figure 1)
provides one avenue for research, there is significant opportunity for testing the effects of
the quantitative vs. qualitative divide in a variety of different contexts and on a variety of

different outcomes. For example, although traditional research has demonstrated the

48

value of cumulative experience quantity on economic outcomes, little has been done on
the effects of qualitatively diverse experience on such outcomes. Alternatively,
evaluating the effects of different experiences on performance consistency or the ability
to create breakthroughs represents another direction. Further, much has been made of
ﬁrrn’s efforts at creating balance between exploration and exploitation, and empirically
assessing the role of different types of experience in promoting such ambidextrous
outcomes could also produce new findings (Gupta, Smith, & Shalley, 2006).

In conclusion, the goal of this research was not only to broaden our conceptual
understanding of organizational experience, learning, and radical innovation but also to
stimulate new research into the effects of experience diversity in organizations. By
providing a new framework for exploring such experience, new theory for connecting
such experience to organizational action and a new model of experience, learning, and
radical innovation, future research will be better equipped to push forward into new

territory.

49

CHAPTER 2

EXPERIENTIAL LEARNING AND KNOWLEDGE CREATION IN THE
PHARMACEUTICAL INDUSTRY

OVERVIEW
This research explores the inﬂuence of different types of organizational experience on
knowledge creation in the pharmaceutical industry. Although the idea that organizations
learn from experience has been well established, to a large extent research on
organizational learning has focused on a single type of organizational experience — how
organizations learn from similar experiences over time. Extending beyond this traditional
approach, in this research I explore how different types of organizational experience
(experiential depth, breadth, and diversity) inﬂuence the rate and impact of knowledge
creation in the pharmaceutical industry. Results from 20 years of data provide support for
the independent and divergent inﬂuences of experience depth and breadth on the rate of
new knowledge creation and the impact of such knowledge. In addition, several forms of
experiential diversity are found to help promote both the rate and impact of knowledge
creation in the pharmaceutical industry, though these effects are often found to have

diminishing inﬂuence at high levels of diversity.

50

Beginning with research on airplane manufacturing over 80 years ago, (Rohrback,
1927; Wright, 1936) and continuing on in force through the post-war economics literature
(Alchian, 1963; Asher, 1961; Garg & Milliman, 1961), the economic benefits to
organizations that build up similar experiences over time have been demonstrated to be
quite robust. In recent years new industries and strategic contexts have been explored, as
diverse as nuclear power generation and pizza-making (Zimmerman, 1982; Darr, Argote,
& Epple, 1995). Nonetheless, over time there has been a consistent focus on the beneﬁts
to organizations of repetitive experience over time.

Although organizations certainly can learn from similar experiences, this
prevailing perspective raises the obvious question of whether or not all experiences are
comparable. Recognizing this potential for qualitative diversity in experience, recently
several researchers have begun to probe the question of how organizations might learn
from different types of experiences. For example, Melissa Schilling and co-authors
examined the efﬁciency benefits of learning by doing something else (Schilling, Vidal,
Ployhart, & Marangoni, 2003). Continuing this trend, Nerkar and Roberts (2004)
explored the effects of technical versus market experience on R&D, and several others
have begun to examine broader conceptions of experience in other contexts (Haleblian &
Finkelstein, 1999, Baum, Li, & Usher, 1999; King & Tucci, 2000). In addition, research
outside of the traditional organizational learning literature in industrial / organizational
psychology has developed new complex models of job experience that integrate
quantitative and qualitative facets of such experience (Quinones, Ford, & Teachout,
1995; Tesluk & Jacobs, 1998). Together, this trend in research provides much of the

foundation needed to push forward our understanding of organizational learning to the

51

differential effects of different types of experience on important firm outcomes, toward a
broader theoretical view of experiential learning.

Building on these ideas. in this study I explore the divergent inﬂuences of
different types of organizational experience on knowledge creation in organizations.
Despite a strong foundation of research on organizational learning, and a growing stream
of research on knowledge creation in organizations, the link between different types of
organizational experience and different knowledge creation outcomes is not well
understood (Grant, 1996; Nonaka & Takeuchi, 1995). Further, extant theory provides
potentially different paths for understanding the nature of these links.

With these issues at the focus, the structure of this paper is designed to address
these questions in three ways. First, I examine the inﬂuence of experience depth on
knowledge creation, building from the traditional literature but hypothesizing mixed
positive and negative inﬂuences on the rate and impact of knowledge creation endeavors.
Second, I examine the inﬂuence of experience breadth on knowledge creation, but this
time hypothesizing opposite positive and negative inﬂuences on the rate and impact of
knowledge creation efforts. Finally, I propose that organizations can promote both the
rate and impact of knowledge creation by fostering diversity in the experiences in their
workforce. Together, these three approaches explore the divergent inﬂuences of different
types of experience on knowledge creation.

To evaluate the role of different types of experience in helping organizations to
create new knowledge, this study investigates experience through the analysis of the US.
patenting activities of pharmaceutical ﬁrms from 1980-1999. By focusing on the context

of intellectual property protection through patenting. I am able to evaluate the degree to

52

which organizations with different types and levels of experience in their R&D

workforces are able to create new knowledge.

THEORY AND HYPOTHESES
Experience and Knowledge Creation

In recent years, knowledge-based approaches to the study of organizations have
increased dramatically (Grant, 1996; Ichijo & Nonaka, 2006; Spender, 1996). Although
this new work holds great promise for organizations, particularly given the rise of a
service economy that puts a premium on the value of knowledge, for many organizations
the greatest promise lies in increased understanding of knowledge creation (Nonaka,
1994). Knowledge creation holds a prominent place in knowledge-based research (von
Krogh, Ichizo, & Nonaka, 2000). in large part because it is often through knowledge
creation that firms are able to create proprietary intellectual property that has the
characteristics needed to provide sustainable competitive advantage (Barney, 1991;
Peteraf, 1993).

Although experience has been described as a key antecedent to effective
knowledge creation, the theoretical links between experience and knowledge creation
have not been well defined. Often it seems as though authors, generally focusing on other
issues in the study of knowledge creation, use implicit theoretical ideas and heuristics to
describe the experiential antecedents of knowledge creation. For example, Nonaka and
Takeuchi (1995) describe experience as an important foundation that underlies
knowledge creation efforts, particularly as experience is a root for ‘creating knowledge in

practice.’ In another example from the knowledge creation literature, von Krogh, Ichizo,

53

and Nonaka (2000) specifically position tacit knowledge as the primary precursor to
knowledge creation, arguing that enabling knowledge creation requires that organization
take steps to unlock such tacit knowledge. This implicitly highlights the role of
experience in knowledge creation, as tacit knowledge is largely gained through personal
experience (Polanyi, 1967). Research on exploration and exploitation, two speciﬁc
approaches to knowledge creation endeavors, has also highlighted the role of experience
in knowledge creation. For example, Holmqvist (2004) builds on March (1991) to outline
the critical role of experience in developing exploration and exploitation capabilities in
organizations, arguing that experiential learning processes play an important role in
exploitation and exploration endeavors within the organizations.

Despite all of this attention to experience as an antecedent to knowledge creation
this research has been surprisingly limited in exploring how different types of experience
relate to the knowledge creation effort. Reﬂecting on this lack of research, Nonaka and
colleagues stated in a recent review article on the future of knowledge creation research
that a key remaining issue is to understand more about the diverse origins of knowledge
creation in organizations (Nonaka, von Krogh, & Voelpel, 2006: 1196). Responding to
this need, I aim to build from existing learning theories to show how different types of
experience (including experience depth, breadth, and diversity) are independent and
differentially inﬂuence knowledge creation outcomes in organizations. Further, as
‘knowledge creation’ is a broad term that encompasses all of an organizations knowledge
creating activities, I will focus on how these different types of experience inﬂuence two
different dimensions of the knowledge creation effort: the rate of knowledge creation

(often studied in the context of exploitation) and the impact of knowledge creation (often

54

studied in the context of exploration). Breaking knowledge creation out into these two
different dimensions helps to highlight the different goals organizations have for
knowledge creation in their own ﬁrms. Although ﬁrms can certainly benefit from both
high rates of knowledge creation as well as highly impactful knowledge creation efforts,
these two outcomes underscore the potentially very different aims of different
organizational approaches to knowledge creation. Further, as will be described below,
extant theory provides potentially different directions for the inﬂuence of different types
of experience on these two knowledge creation outcomes, and provides a framework for

exploring the divergent inﬂuences of experience in organizations.

Experience Depth and Knowledge Creation

To begin with, depth in experience — gained from repeated encounters with many
similar situations over time — is argued to positively inﬂuence the rate of knowledge
creation in several ways. First, experience depth allows organizations to develop efﬁcient
practices, search short-cuts, and helpful heuristics in a particular line of work. Further, as
these efﬁciency improvements and short-cuts are accumulated throughout the
organization, they can be institutionalized into standardized procedures (Pugh, Hickson,
Hinings, & Turner, 1968), helping to increase the efﬁciency of future work. An
organization that is new to a technical area needs to experiment to try and find the best
ways to do things. In contrast, an organizational workforce with extensive experience can
simply focus on efﬁcient execution. This focus helps to increase the rate of knowledge
creation by reducing distractions, helping to create an organizational competency for

efﬁcient execution. Second, increasing levels of depth in workforce experience facilitates

55

 

the development of centralized information and search centers, knowledge management
systems, and knowledge repositories (Davenport & Prusak, 1998). In this sense, an
organization that has a long set of similar experiences is likely to develop centralized
systems to make gained knowledge widely available, providing a convenient and
proximal search environment for future work. This helps to increase the efficiency of
knowledge creation and promotes the rate of knowledge creation. Finally, organizational
workforces with extensive knowledge in an area can improve their rate of knowledge
creation by speeding the turnaround on bad ideas, as they are working in well understood
knowledge landscape, making it possible for design problems to be solved faster
(Fleming & Sorensen, 2004). This increased speed in problem solving will help support a
more rapid schedule of producing new knowledge. Together, all of these ideas build on
the traditional leaming-by-doing literature (Argote, 1999) to hypothesize:

Hypothesis 1 a: The overall level of depth in the experience of an organization 's

research workforce will be positively related to the rate of knowledge creation by

the ﬁrm.

However, despite the promise of efficiency from the traditional learning curve
literature, there is also a potential ‘dark side’ to experience depth that has been less well
explored. Specifically, deep workforce experience may hinder the creation of high-impact
knowledge. This is hypothesized for several reasons. First, building on the work of
Leonard-Barton (1992) and Dougherty and Heller (1994), increasing levels of experience
depth promotes the development of core rigidities over time. In this sense, core rigidities
support a dominant paradigm within organizations that can act to stiﬂe new ideas.

Building on this. increasing depth also supports the rise of a dominant logic in the

56

research workforce and management to support certain projects and reject Others.
Although this type of focus is probably good for the rate of knowledge creation (Hla) it
creates resistance to out-of—the-box thinking that helps to create breakthroughs (Leonard-
Barton, 1992). In the pharmaceutical domain, scientists who work narrowly within these
organizationally-supported domains would be less likely to be exposed to new ideas or
alternative cognitive structures and thus less likely to ﬁnd or pursue novel ideas. Over
time, organizational systems of evaluation and control come to support the dominant
paradigm, and those who support the dominant paradigm gain elevated status in the ﬁrm
(Leonard-Barton, 1992). Together, these factors support the creation of myopias to
outside thinking (Levinthal & March, 1993) and reduce the potential for creating high-
impact knowledge.

These ideas support the following hypothesis, that experience that helps a firm
succeed in one dimension of knowledge creation (the rate of knowledge creation) may
directly restrict their ability to achieve success on another dimension (the impact of new

knowledge).

Hypothesis 1b: The overall level of depth in the experience of an organization ’s
research workforce will be negatively related to the creation of high-impact
knowledge by the ﬁrm.
Experience Breadth and Knowledge Creation
On the other hand, breadth of organizational experience — gained from encounters
with many different situations over time — is argued to ﬂip these relations such that it

positively inﬂuences the impact of new knowledge but negatively inﬂuences the rate of

new knowledge creation. Speciﬁcally, broad workforce experience is proposed to hinder

57

the rate of new knowledge creation for several reasons. First, with wider breadth of
workforce experience a tendency can develop to explore a wide and varied set of
information sources to find ideal solutions to problems. Although this helps to promote
the impact of such solutions, delays can slow the pace of innovation. The time spent
searching through a diverse set of alternatives directly impedes the speed and efficiency
of knowledge creation efforts, leading to a slower rate of new innovation. Second,
experience with a diverse set of information sources may make it difﬁcult for researchers
to focus narrowly on a single set of ideas or narrow projects. Supporting this idea,
research on information processing has shown that in teams where efficiency is expected,
providing access to diverse information sources impedes the efﬁciency of team
performance due to information overload (Katz & Tushman, 1979; Keller & Staelin,
1987). Finally, again building on work of Leonard-Barton (1992), as was argued
previously, core rigidities can develop over time to support a dominant paradigm in an
organization. In this case scientists and managers become engrained in routines of wide
search rather than more focused efforts, creating rigidities toward breadth that can be
problematic. In such cultures breadth of search, creativity, and novelty can become the
standard and the norm (rather than efﬁciency) and any solution brought to the table will
be evaluated for its novelty and origin. Well known examples of such creativity focused
cultures are those of Apple computers, Virgin group, 3M, and IDEO product design
(McGregor, 2008) where the focuses on creativity can itself be a force to overcome to get

products to market.

58

These ideas support the following hypothesis, that experience which helps a ﬁrm
succeed in one dimension of knowledge creation (the impact of knowledge) may restrict
their ability to achieve success on another dimension (the rate of knowledge creation).

Hypothesis 2a: The overall level of breadth in the experience of an organization 's

research workforce will be negatively related to the rate of knowledge creation by

the ﬁrm.

However, in addition to this negative inﬂuence of experience breadth on rate,
once again I propose that there is also a potential tradeoff associated with this experience.
Although breadth is argued to negatively inﬂuence the rate of new knowledge creation, I
hypothesize a positive relationship with the impact of such created knowledge. This
opposite relation is proposed for several reasons. First, with broad workforce experience
available an opportunity is presented for the organization to explore and grow familiar
with a varied set of information sources. Such diverse sources of knowledge inside the
firm then aid future problem solving to make a wider variety of familiar alternatives
available to ﬁnd better solutions (Grant, 1996) and potentially create high impact new
knowledge. Second, a diverse set of workforce experiences provides the organization
with alternative cognitive models (Pitcher & Smith, 2001) to use in developing new
products and services. These unique cognitive models provide a variety of distinct lenses
through which to consider new problems, and may lead to alternative framings of
problems within the ﬁrm. As these different framings can lead to new and novel
combinations of ideas (Henderson & Clark, 1990), high impact knowledge is more likely
to emerge. F inally, because the organization has a wide base of experience, this allows

the organization to develop skills in the use of different types of research tools (e. g.

59

databases, patent libraries, physical lab equipment, etc.) and methods in-house. This helps
to ameliorate the problems of ‘spatial myopia,’ (Levinthal & March, 1993), the tendency
for individuals to rely on information and resources available near them. As organizations
with wide experience sets have diverse skills and knowledge sets located in a socially
proximal environment (in-house), this makes it easier and more likely for individuals to
engage in wide search for novel solutions to problems.

These ideas continue to build on the trend in the recent experience literature to
explore new and broader conceptualizations of experience (Nerkar & Roberts, 2004;
Schilling, et. al, 2003) and support the following hypothesis:

Hypothesis 2b: The overall level of breadth in the experience of an organization 's

research workforce will be positively related to the creation of high-impact

knowledge by the ﬁrm.
Experiential Diversity and Knowledge Creation

Up to this point, the proposed hypotheses have painted a somewhat bleak picture
of knowledge creation for organizations hoping to balance the rate and impact of their
research and development efforts. Building a research workforce with strong experience
depth was hypothesized to promote the rate of new knowledge creation but came at the
cost of hindering the impact of such knowledge. On the ﬂip side, experience breadth was
argued to do the opposite, promoting impact but hindering the rate of new knowledge
creation. With each of these paths, organizations would be forced to make a choice
between rate and impact.

In contrast, this section of the paper suggests that organizations may be able to

promote both the rate and impact of their knowledge creation efforts by fostering

60

experiential diversity in their workforce. Rather than choosing a single direction in
knowledge creation, this approach suggests that organizations could have a high rate of
knowledge creation and yet still create high-impact knowledge.

Experiential diversity occurs when individuals with high and low levels of
experience work together toward a common goal. In the context of this research, this
could happen either by mixing individuals with deep experience together with those who
are new to a knowledge domain or by mixing those who have broad experience with
others who have very narrow experience. For example, one organization might have a
group of scientists working together who have each worked in their ﬁeld for over twenty
years (homogenous high depth). Another organization might have a more diverse group
of a few scientists with twenty years of experience and a few with less than ﬁve years of
experience (high experiential diversity). From a breadth perspective, a diverse group
might be a collection where some scientists have worked in ten different therapeutic
domains mixed together with other scientists who have only worked in one or two
domains (high breadth diversity).

These types of experiential diversity are hypothesized to promote both the rate
and impact of knowledge creation in two primary ways — by facilitating knowledge
sharing and by encouraging task-based cognitive conﬂict. First, mixing individuals
together who have varying levels or types of experience is posited to encourage
knowledge sharing. With significant diversity in different employee’s experiences, an
opportunity emerges for researchers with deeper or broader experiences to share their
gained knowledge (both accumulated insights and content from other domains) with their

less experienced colleagues. This stands in contrast to homogenous workforces where

61

such knowledge might not be shared because less task-based dialogue occurs between
similarly experienced individuals as compared to the dialogue that is common between
experienced and inexperienced individuals (Sundgren, Selart, Ingelgard, & Bengtson,
2005). In addition, this diversity in experience helps to mix old and new ideas in a way
that has been demonstrated to aid the development of creative new knowledge, through
what Leonard-Barton called “creative abrasion” (Leonard-Barton, 1995). This empirical
research on the ‘age’ of technical knowledge has suggested that key insights can be
created through the bridging of ‘old’ ideas and ‘new’ ideas (Katila, 2002; Nerkar, 2003).
In this sense, by creating diversity in experience, the ‘old guard’ of experienced
professionals (some with signiﬁcantly outdated knowledge) can interact with new college
graduates (potentially trained in new knowledge paradigms) to create cognitive friction
that helps to promote the development of new and potentially high-impact ideas.

Second, mixing individuals together who have different levels or types of
experience can also promote task-based cognitive conﬂict (J ehn, 1995; Jehn, et. al, 1999;
Pelled, et. al, 1999). In this context, task-based cognitive conﬂict occurs when individuals
working together disagree on how to best approach or solve a problem. Experiential
diversity makes such conﬂicts likely as people often rely on their previous experiences to
guide them in solving problems. When people draw from signiﬁcantly different
experiences, conﬂicting ideas are likely to emerge. Although some types of diversity can
be problematic for groups by creating relationship-based conﬂict (J ehn 1995; 1997),
cognitive conﬂict of ideas has been often shown to productively inﬂuence group
outcomes (Amason, I996; Pelled, Eisenhardt, & Xin, 1999). However. recent meta-

analytic analysis of the conﬂict literature has noted that despite this evidence, task-based

conﬂict it not always positive, and additional research is needed to further understand the
context in which such cognitive conﬂict helps teams perform (De Dreu & Weingart,
2003). The positive ﬁndings for task conﬂict suggest that “teams beneﬁt from task
conﬂict when they cultivate an environment that is open and tolerant of diverse
viewpoints and work with cooperative norms preventing those disagreements from being
misinterpreted as personal attacks” (De Dreu & Weingart, 2003: 747). In addition,
research on top management teams has suggested that managers can do well to cultivate
information-based diversity (Simons, Pelled, & Smith, 1999), particularly when such
diversity culminates in ‘having a good ﬁght’ (Eisenhardt, Kahwajy, & Bourgeois, 1997)
between different ways of thinking. Related research on devil’s advocacy, where people
take purposefully divisive positions on topics, also highlights the importance of diversity
in opinions within organizations (Schwenk, 1990), as such strategies improve strategic
decision making (Schweiger, Sandberg, & Regan, 1986). When this type of conﬂict
occurs it promotes increased discourse as individuals defend their different positions. and
such increased communication can get more ideas out in the open. With such expanded
discussion of task-based ideas, creative breakthroughs become more likely, as creativity
is aided by such broad and open dialogues (Wanous & Youtz, 1986).

Through these two processes both the rate and impact of knowledge creation
endeavors can be enhanced. The impact of knowledge creation is helped as the increased
knowledge sharing and cognitive conﬂict facilitate the ﬂow of new ideas for products
that often combine distant sources of experience (technologically, organizationally, and
temporally) to create potentially novel new knowledge. Experiential diversity can help to

promote the rate of knowledge creation in two ways. Such diversity helps to make

63

potential product extensions more likely to be recognized and discussed (identification of
new potential markets for existing products / knowledge), creating an enhanced ability
for the ﬁrm to recognize new commercialization opportunities. As a lack of
commercialization opportunities is often cited as a key impediment to the rate of
innovation (Gans & Stern, 2003; Shane, 2001), factors that help to overcome this obstacle
should promote a higher rate of knowledge creation. In addition, diversity can help to
improve the rate of creation by providing a forum and impetus for the sharing of best
practices for efficient work (von Krogh, Nonaka, & Aben, 2001; Verona & Ravasi, 2003)
as well as benchmarking for continuous efﬁciency improvements (Winter, 1994). In
summary, these arguments suggest that it is the presence of different levels of knowledge,
not the overall quantity, which helps organizations to achieve both high rate and impact
in their knowledge creation.

However, building on these ideas, it is not clear that the relation between diversity
and knowledge creation would be linear or uniformly positive. In fact, it may be possible
that this proposed relation between experience-based workforce diversity and knowledge
creation could be subject to diminishing returns, or even negative returns, at high levels
of diversity. Specifically, for many of the same reasons diversity was argued to be good
for knowledge creation, such as promoting knowledge-sharing and cognitive conﬂict, at
the extreme such inﬂuences could be harmful. Too much of a good thing can become
counterproductive. For example, when absolutely no agreement exists (everyone is an
island), diversity can be counter-productive as individuals may not be able to ﬁnd
common ground to communicate effectively with each other. Further, although a

moderate degree of task-based conﬂict can be good (J ehn, et. al, 1999), too much conﬂict

64

 

may overwhelm the realized benefits and conﬂict can become the center of social
interaction (Van de Vliert & De Dreu, 1994). As a result. these hypotheses will posit that
experience-based diversity will exhibit an inverted-U shaped relation with knowledge
creation outcomes.

H3a: The diversity of experience reﬂected in an organization’s research

workforce will exhibit an inverse-U shaped inﬂuence on the rate of knowledge

creation by the ﬁrm.

H3b: The diversity of experience reﬂected in an organization ’s research

workforce will exhibit an inverse-U shaped inﬂuence on the creation of high-

impact knowledge by the ﬁrm.

METHODS

Data and Setting

The value of the rate and impact of knowledge creation is likely to vary by
industry such that in some industries these outcomes will have much less value for
performance than in other cases. As such, to evaluate these hypotheses empirically it was
important to choose an industry setting where both rate and impact are likely to be
valuable. Although several potential industries that meet this criterion were considered,
research on the pharmaceutical industry (SIC 283: NAICS 3254) provides signiﬁcant
support for the value of both dimensions of knowledge creation in the production of
commercial drugs (Bierly & Chakrabarti, 1996; Rothaermel & Deeds, 2004). Thus this
study investigates the workforce experience of 167 unique ﬁrms in the pharmaceutical
industry, which included all of the ﬁrms in this NAICS code for which data were
available. The time period for study are the years 1980-1999, creating a pooled

longitudinal and cross-sectional panel dataset. This time period was selected because

65

 

 

1980 was the first year in which full citation information is reliably available for patents,
and 1999 because this is the latest year full information is available from the National

Bureau of Economic Research (Hall, J affe, & Tratjenberg, 2001).

Measures

Patent data were used to create the experience variables. This approach is
consistent with research in recent years that has examined innovation in organizations
(Ahuja & Katila, 2001; Rosenkopf & Nerkar, 2001; Sorensen, et. al, 2006) and provides
an in—depth look at the experience of individuals who engage in research within
organizations. As noted by Sorensen and colleagues (2006), patent data offer something
of a “fossil record for following the ﬂow of knowledge and the build up of experience,
providing a lasting reﬂection of ephemeral interactions.” Although the use of patent data
has some signiﬁcant limitations (see Agarwal & Henderson, 2005; Mowery, Oxley, &
Silverman, 1996), including the exclusion of many potentially relevant non-patent-related
experience and members of the research workforce, the clear trail of experience reﬂected
in these archival data provides a detailed portrait into the experience history of individual
scientists in an organization’s research workforce. Further, the use of patent data allows
our analyses to focus on some of the most important knowledge in these ﬁrms (that
which the organization thought was important enough to patent) and provides a method
of consistently measuring knowledge creation between companies and over time. In
many ways patents represent one of the most concrete measures of the abstract

organizational knowledge construct. As such, these data provide a deep and objective

66

source for the examination of depth, breadth, and diversity in the experience of research

scientists in organizational workforces.

Independent Variables

The calculation of the workforce experience and experience diversity variables
was completed in several steps, starting at the individual scientist level and building up
through aggregation to understand the composition and diversity of an organization’s
patenting research workforce. First, the individual patenting histories of every scientist
who was named as a contributor on a US. patent from 1960-1999 were calculated to
determine the depth and breadth of each individual scientist’s experience on a patent-by-
patent running basis, regardless of the speciﬁc organization for which they worked or
industry in which they were working. Second, for each of an organization’s patents in
each year from 1980-1999 (the years in which adequate patent information were
available) the average depth and breadth in experience of each team of researchers named
on these patents were calculated. These patent-level experience scores were then
aggregated up to the organizational level to include all of an organization’s patents in the
focal year. Next, the diversity of each organization’s workforce was assessed similarly,
by calculating the standard deviation in each of the individual-level experience variables
(depth and breadth) among scientists named on each patent. These patent—level diversity
scores were then aggregated up to the organizational level (averaging the diversity among
all of the organization’s patents in the focal year) to give an average diversity score

across all patent teams in the organization in that yearz. The selection of standard

 

'1 . . . . . . .
‘ An alternative to this 3-stage process of scientist experience accumulation. patent level aggregation, and
then organizational level aggregation is to simply pool all named scientists on the organization’s patents in

67

deviation as the measure for diversity in this context was a careful decision. This decision
was reached by following the theoretical guidance provided by Harrison and Klein
(2007), in which the type of diversity of interest in this research is called ‘separation’-
based diversity. Harrison and Klein (2007) argue that the most appropriate and well-
accepted measurement of separation-based diversity is the measurement of standard
deviation in the focal variable. As such, standard deviation was used in the measures
below to represent diversity.

Speciﬁc measures for experience depth and breadth are outlined below. To
provide a broad and rigorous test of proposed theory, two measures were calculated for
each type of experience of interest and included both separately and together in various

econometric models.

Dept/1 of Workforce Experience

Number of Patents. The first approach used in the experiential learning literature (Argote,
1999) as a measure of the depth of experience is volume, in this case measuring the total
volume of cumulative experience an individual has accumulated in their previous work.
This is measured as the total number of patents on which an individual has ever been

- 3
named as a contributor.

 

a given year and assess the depth. breadth. and diversity of their experiences as a collective pool. To
determine if the analyses were sensitive to using a pooled approach rather than the 3-stage aggregation
method. I ran models using both approaches. The pattern of results was largely the same regardless of the
method of aggregation. I chose to present the 3-stage approach for the primary analyses. as this approach
more firmly places the measurement emphasis on teams of researchers working together, which is
particularly important given my theoretical arguments regarding experiential diversity. in which increased
discourse and knowledge sharing require ‘creative abrasion’ (Leonard-Barton. 1992) to occur between
individuals with different experiences actually working together.

3 An alternative approach to the assessment of experience depth would be the measurement of an
individual’s previous work only in the exact same technical domain as a current project. This approach was
rejected in favor of a more inclusive approach that measured an individual’s work on any previous patents.

68

Years of Experience. The second approach used in the experiential learning literature
(Argote, 1999) as a measure of the depth of experience is time, in this case measuring the
total amount of time an individual has been actively involved in patentable research. This
is measured as the total number of years an individual has been actively working on

patents up to each given period.4

Breadth of Workforce Experience

Technological Breadth. The first measure of breadth in an individual’s experience is a
measure of the total range of an individual’s technological experience. This is measured
by a count of the number of different technical subclasses in which the individual has
been granted patents in the individual’s career. In this sense, if a researcher had been
granted patents in 20 different pharmaceutical domains, this would represent that
individual’s technological breadth.

Organizational Breadth. The second measure of breadth in an individual’s experience is

another measure of the range of an individual’s experience. This is measured by a count

 

regardless of exact technical class. to give a more complete picture of each scientist’s true depth of patent
research experience. Nonetheless. depth variables were also calculated using this more restrictive approach
to depth measurement. Models that used these more restrictive depth measures followed the same pattern of
results as the primary analyses. but were generally weaker in terms of signiﬁcance. likely due to the
omission of many relevant prior experiences that were not exactly in the same technical class as a given
project.

4 Again. an alternative approach to the measurement of this variable is possible. In this case. ‘years of
experience’ was assessed as a count of the number of previous years in which a scientist actually ﬁled a
patent. Although 20 years may have passed between a scientist’s first patent and the current year. if they
only were named on patents in 10 of the intervening years. their ‘years depth’ score would be ID. The
alternative is to count all years that have passed since a scientist was named as a contributor on their ﬁrst
patent (20 years in this example). The count of active years rather than all years was selected to give a more
focused look at the active experience histories of scientists. and to avoid errors of measurement that would
give a scientist 30 years of experience credit in the hypothetical situation where a scientist had a single
patent 30 years ago and none until the current period. However. to assess the influence of this measurement
decision on results. both approaches were assessed empirically. Again. models that used the less-precise
measure that just counted years from a first patent were generally lower in level and significance than the
more focused ‘active years’ approach. likely due to the inclusion of noise in measurement of the underlying
construct.

69

of the number Of different outside organizations for which the focal scientist had worked
and created patents in the past, prior to the focal year. In other words, if an individual
from Merck had only created patents for Merck in their career, this would indicate low
organizational breadth with a value of 1. If, however, the individual had previously
submitted patents while working at 3 or 4 unique organizations, this would indicate high
organizational breadth. These two approaches to breadth in patenting (technical and
organizational) build on the work of Rosenkopf and Nerkar (2001) who argued that both

of these types of breadth are important to understanding innovation in ﬁrms.

Diversity of Workforce Experience

In addition to the average level of workforce experience, to measure the diversity of
experience in each organization, the standard deviation of each of the four measures
above was also calculated. As noted above, standard deviation follows Harrison and
Klein’s (2007) approach to separation-type diversity and helps to capture the degree to
which each workforce is characterized by signiﬁcant differences in experience between
research scientists. To evaluate the inverse-U shaped relation hypothesized in the effects
of experience-based diversity on the development of knowledge by organizations, each of

the diversity terms were squared.5

 

5 Issues with collinearity in multiple regression become particularly salient when models are evaluated that
include both theoretical variables and their squared terms. Although there are different approaches to the
evaluation of such models (Gordon, 1968). including centering higher order terms prior to including them
in analyses. such transformative techniques have been criticized (Tate, 1984). To address the potential for
bias in such models. but still represent the important effects of these terms in such models (Friedrich.
1982). several steps were taken and are described on pages 71-73.

70

Dependent Variables
Rate of Knowledge Creation
Number of New Patents Granted. The ﬁrst of the dependent variables is the measure of
the number Of patents received by an organization in a given year, assessing the rate at

which new patentable knowledge is created by the focal ﬁrm.

Impact of Knowledge Creation

Number of High Impact Patents. Several measures of the impact of knowledge creation
have been proposed through the years (Dahlin & Behrens, 2005). For this research, the
focus was simply to assess the ability of an organization to create new knowledge that
signiﬁcantly impacted future research in the ﬁeld. With this goal, the most widely used
measures of impact in the literature on knowledge creation in organizational research are
those that measure the impact of a new patent on the development of future patents,
measured by the citation counts of such patents (Ahuja, 2000). To measure this variable,
each patent is compared to other patents filed in the same year’s cohort. In each cohort
year, the average number of citations received by patents is calculated, and then two
measures are proposed to indicate whether a particular patent had a high impact. These
two levels of impact are the calculation of patents that received citations at a rate 1 or 2
standard deviations above the mean number of citations in a given year. Together these
two levels provide indications that a patent had a signiﬁcant impact on the field, but at
varying levels of impact. In addition, by making each assessment of impact relative to
other patents in a speciﬁc year’s cohort, the issue of non-comparability of impact from

one year to another across 20 years is mitigated, as the degree to which each year’s

71

patents are highly impactful is evaluated relative to other patents originating in that year.
Together, these measures of widely cited patents are a strong indication of an

organization’s ability to create high impact knowledge.

Control Variables

By limiting this study to the investigation of the pharmaceutical industry,
industry-speciﬁc alternative explanations for the results are largely controlled. This was
perhaps the most important control factor considered, as several of these industry-
relevant factors include the inﬂuence of industry-level technical complexity, market
competition, and legal / regulatory restrictions on business practice. In addition, to further
control for the possible inﬂuence of other firm-specific factors, several additional control
variables are included in these analyses.
Firm R&D Expenditures. Firm research and development investments are an important
determinant of a ﬁrm’s patenting activities. As such, and to eliminate the possibility of
this factor confounding or providing an alternative explanation for achieved results, the
dollar amount of each ﬁrm’s total research and development expenditures each year is
included as a control.
Size of Parenting Research Workforce. To control for the differences in the scope of
patenting research being carried out by these firms, research workforce size will also be
included as a control variable, measured as the number of unique individuals that were

named as part of the patenting activities of the ﬁrm each year.6

 

° Again, as noted earlier. the evaluation of only researchers named on patents is an imperfect approach to
studying knowledge creation in organizations as it excludes 'unproductive' research personnel who did not
achieve patentable knowledge in their work. and other scientists involved in knowledge-creating work in
the organization that did not have a natural termination in a patent. However. this does very ﬁrmly place

72

Pharmaceutical Sub-Domain. Building on the idea that subtle differences in knowledge
creation could potentially exist between different pharmaceutical industry sub-domains,
indicator variables for each subclass were included as control variables. These subclasses
are described in Appendix A.

Fixed Firm Effects. Fixed ﬁrm effects were necessary to control for unobserved
heterogeneity between ﬁrms not captured in other controls. This choice was supported by
the Hausman test, which showed significant differences between the fixed and random

effects estimations. indicating that the ﬁxed effects were appropriate for this data.

Econometric Analyses

To evaluate the proposed hypotheses, several different estimation procedures
were necessary. Models with the rate of knowledge creation as the dependent variable
were evaluated using fixed effect negative binomial panel regression, a typical analysis
for such data (Hausman, Hall, & Griliches. 1984; Fleming & Sorensen, 2004) in the case
of count-based dependent variables with demonstrated high levels of overdispersion
(based on test of the overdispersion parameter alpha). This overdispersion led to the need
to use negative binomial estimation rather than Poisson estimation, as Poisson estimation
makes assumptions regarding mean and variance that are not viable in overdispersed
data. This complex regressive technique is necessary due to the fact ordinary least

squares analysis (OLS) can give biased and inefﬁcient estimates when analyzing

 

the focus on some of the most critical knowledge and research personnel in an organization (knowledge
that is so important that the organization needs to patent it to secure legal property-right protections), and
thus provides a well-placed window through which to view experience and knowledge creation in
organizations.

73

longitudinal panel data, as the year-to-year observations of specific firms are not fully
independent, violating OLS assumptions regarding the residuals in such analyses.

However, models with knowledge creation impact as the dependent variable had
an additional problem that required further attention. As you can see from Table 3, a fair
percentage of ﬁrms never created a high impact patent, particularly a 2 standard-
deviation high impact patent, leading to a large number of zeros in the dependent variable
for the knowledge impact models. Standard negative binomial regression is susceptible to
bias in the case of excess zeros (Cameron & Trivedi, 1998). The solution is to use zero-
inﬂated negative binomial regression, which speciﬁcally accounts for the excess zeros, as
well as overdispersion and unobserved heterogeneity in the data (Long, 1997). The test to
determine deﬁnitively if zero—inﬂated negative binomial regression is statistically
superior to other approaches in a dataset is the Vuong test (Vuong, 1989; Greene, 1994).
Vuong tests in the statistical package Stata were significant at a 0.001 level for all
knowledge impact models (Vuong test statistics ranged between 3.04 to 6.10), conﬁrming
that zero-inﬂated negative binomial regression was the appropriate choice for these data.
Thus, Tables 3 and 4 report the zero inﬂated negative binomial regression results for
predicting the number of high-impact patents created by an organization.

To assess the potential for multicollinearity in the analyses, several diagnostic and
analytical steps were taken. First, diagnostic tests using the ‘colldiag’ function and
variance inﬂation factors were evaluated in Stata to assess the presence of
multicollinearity. The proposed analytical models did demonstrate significant levels of
multicollinearity in fully saturated models that included multiple measures of the same

theoretical constructs. This was not surprising given the high bivariate correlations

74

between some variables as shown in Table 3, and the fact that these measures were
specifically selected to provide multiple measures of the same theoretical constructs.
However, less saturated models that included only subsets of variables were free from
problematic levels of multicollinearity. As a result, to address these concerns, but also to
fully represent the data and proposed hypotheses, separate econometric models are
reported that contain both fully inclusive as well as selective subsets of the measures. For
example, in the Table 4 reporting of the relations between experience and knowledge
creation rate. models I, 3, 5, and 7 assess the hypotheses with only one measure of each
type of experience included. In addition, models 2, 4, 6, and 8 include these measures
together with their diversity and diversity squared terms (for evaluating Hypothesis 3).
Model 9 then evaluates all measures of all constructs together in one fully saturated
model. This pattern is repeated for the knowledge creation impact dependent variables in
Tables 5 and 6.

By reporting analytical models that assess these various types of experience
together and independently I am able to evaluate the unique contributions of each
different type of experience on the various knowledge creation outcomes. This is
important, as it helps to answer questions about how these different types of experience
come together to both jointly and independently inﬂuence knowledge creation in

organizations.

75

£2.20 .m.D E 265:: E 3.53.. as Benz—Oceaxm Odom :5”. _
.5. v a Enchwﬁ 2a 5. SE. 532w 25.3850 .3. v m .58....ch us me. SE 558w 223.250

 

R. 2. c... 2. t. 2.. E. 5.. mm. 1. m... 8.- S.- 8: 3.- 3.- 5m S... 8.2% 55.36 3:23:35 .2
-- a. a... R. R. 2. on. .m. on mm. 2. 3.- no. 5. 8. 8. 8.2 .c... 8.33.3226.3325; .c.
on 3. 3.. E. i. 8.- 8.- 8.- S- 8: Ed 3d 5.82.53:23.533-2330 .m.
E. on. E :.. 8. 2. .5. z. 2. m: 8.. .9236 ac.w.._o._..§-b.eo>5 .2

- Essa-8:536 2.9:: .o u
2. E. :. 2.. x3.- 2: 8.- .5.- S: N... Na... Mcﬁeem 8.3...er .m.

. - mo_wc_2_coo.osvz_= .o n
2. N... 3. E. 2..- 8. .5. .5. 2.. 4mm m: «5.85 85.8.6 .N.
-- a». so. 3.. 8.- t. 8.- 5.- 8. a; 5+. Bagging 5832...; .:
E. 5. mm. 3. S. .c. we. 2. 3. 2.3 3.3.;- babeu 583.52.... .3.
-- i. o... 2. 2. 8. 2.. 2. mo. 8.. outage-3.5....»-2330 .o
-- 9-. 2. 2. we. 2. a. who N3- 523... 5.8 25...... - ripen. e
-- 5. 2. S. .5. A... m: .56 e8» 2 a - £95 88:85 N
.5. 2. 2. 2. a. m2: 2.... 25.3% a - £95 caeetaxm e
- an. $.- cm. 3.. Q:- .m..m 85.3.63 533.3. ”.5525 .6 aim .m-
-- a... 3.. a. 38 2.3 @2286 93. E.5. ...
-- .. _.

. . 3:623:35 E23. .32:
-- x: mm ex 2 2mm .23....65 owl—5.33:! ._

2 a w A. m a m N _. .Za .9m :32 «SEE;

 

moEEB> 3.2m 563.2. macaw—8.50 I m 635.

76

OBS £5 E 8026:. .0: 03 3565666 :65 :5 3936:. $233 25.6... £252 _
.25. v a. .13.. :o. v a .1. ”mo. v m * 5.. V m ... .mcocmioﬁo wow act... no. u 2 ”2:2

 

 

mom «.9. wow wow wow wow wow wow wow 2
*szoov ***ocm ***mvm tel-mm ***mwm ***Nmm ingom ***©om ".23..va NRESS
***Cm._ ***mm._ ***_m._ .3..ng ***o_._ ***m_._ .I..: _._ “IRE.— ***:._ ESEOU

me: $33.. 35:9- 5??? Ecosnﬂcaweo

* 5.- ***mc.- cuss—om habit _mo_wo_o:563 Stoogm

ism. 333m. 383% 3:83.590 2.. b...£.y:Q

.1 _. ism-m. b.3926 Ramsay—6o.

u«m. .- No: max Eczaﬁcawuo 2.3:: Co # .353m
.331: . time. 3.43. 3.3.2.52 2.3:: .o t 3.335%
ice.- ..tc...- assay... 5.23% 5%.. was»

ac. * . or cuss—om 5.33% 5ro E29. .3ch
$3.... a... _ m. gaze ease as; 5 based
cc. *3. .9336 Ego—O 2.23
mg: *anWO. “weave. ”—30% .«O # <->uQ
cc. ** .0. .31. _o. 3:23 ho a. 3:...Ei...m
“sixes. *xsoc. ﬁasco. ***cc. 0.3.3403. xxxmo. *xaoo. *snnmc. fibe- OBQV—LOB £9533” ho 35
".3." _©.- ** 5.- ”.2. _o.- “.313.- ".3." _C.- ".3." _O.- ".319.- ** 2v.- ** _c.- muszﬁcoaxm Gum“ Etn— wry-«:59
* * * * * * s * * 29.300 mmn_un:m
a Etc—2 w .252 h .232 w .252 m .252 v .252 m .252 N .252 _ .252 033.3.» mac-.9

 

83— :65:95 auto—>55—

 

Aoam c2820 036.365: 3.53% cove-83m .EEoEm o>cmwoz .ootm 63E I a 633.

77

wow wow

.032 0.5 0. 0000.00. .0: 0.0 200.2..000 0.2.. :5 20.00.00. 80.00.00 0020:. 0.0005. _

.00. v m .23.. ".0. v A .2. "m0. v a. ... ”0.. v Q 52.0.3202? wow HE... no. u 2 ”20>.

 

 

 

wow wow wow wow wow wow wow 2
-N
flan. 332K. :15. .I.....»N. ***MN_ term: 02.3110. .135. . ﬁlm: 0.50% 3.5.30.5

m _ .C. 3330.0. ***0m.c- ***mc. . - ***m0.o- .2330. . - ***m...c- 33.00.- 3.35.- 2:20.000

.5... 00. 003:3 b.2030 .0:0.E~.:0w.0

3100.- .30.. 008:3 b.2030 30.90.0532 5000.0

3.. *3..- 5.2030 300.200.0030 3 0.50.5
51.3. .1 .. b.2030 30.3.0052
1.0m: 00. *2..- 20.03.00.000 020.0: .0 0 5000.5

.00. *. .. ***m.. m0.w0.0:_.02 000.0: .0 # 00:0...05m

00. .00.. 0030.00. b.2030 £000 200..

00. 00. 00.01% b.3030 5000 E23 .
*3..- m0: 0.2030 .0000 £00.. 030m. 3 0.50.25
.0: m0: 5.2030 5000 E23
*0. .- $0.. *3 .- £00.. .0 u. .
3.0. .315. 1.0. 2:200 .0 # <E0Q 0.50.205m

Inﬁll... .2335. .1330. .2133. 032...... ***wc. .I....bc. ***..3. 0231.3. 00.0.0103 50.033. .0 0~.m
00. 00. 00. 00. 00. 00. 00. 00. 00. ”0.3.0003”. Q03. FEE 00.2.5.0

* * * * * * * * * 10.2.00 30.09%

0 .0005. w .0002 h .0005. 0 .0005. m .0005. v .0005. m .0005. N .0005. — .0005. 030.00., @0006
. 0000:.— 0000000 300.305—

 

.. .0008. 00.80.“. 0w00.300v.. 2.30m 00.300w0y. .2800... 0>..0w02 020.....-20N I m 0.00..

78

 

 

.033 25 a. 8.5.9:. 5: 25 35.0508 :2: :5 8.50 5:85 333.5 25.2.. 20.5.). _
.25. v a “.33." ”.o. v a "in ”mo. v Q * 5.. v a. .. .mcogtomno wow ”£5... E. .I. 2 “2:2

 

 

wow wow wow wow wow wow wow wow wow 2
-N
“133%. ***mm ***.m ***wm film ***Nv *ﬁam ***mm .13me 3.232 EMS....E«5
$6- 51:..- mmd- 135m..- "wimm..- “W33..- “wimm..- 33m... $33...- .5250
*wm. m .. .5333. 5.336 .aco..a~.:aw5
* 2..- 5.. 332.5 3.33% .3.mc.c=52 5.53%
*3: *co. . - 5333.5 35:35me E ...:.,..§:Q
*nm. *3. b.3035 .ao.wo.o::02
mo: N. . .- *3..- 3053.590 25E: .0 a. SEVEQ
*o.. it.” ***0m. 8.5.2503 2.3:: .5 # $52393
0.. cc. .5333 5.9635 £95 £3»
5.- co. 3:33 big... Egon 523 £35
*mar 5.x- 5.336 £525 23» 2.. 5.25....Q
co. *m . .- b.2035 5ro E23
.m: was we: 33» .o t :5va
”we. iiwo. ***E.. £523 we a. 9:3...SSMN
*3. ...mo. .29. flue. **mo. ”1%... “133%. ***cc. ***3. 09.85.53 £0533. .0 aﬁw
C... a... so. .5. co. cc. 03. 8. cc. meaﬁcoaxm Odd. Eb". €5.59
* * * * * * * * * 22:50 $3.55
a .255. w .353. h .25.). e .253. m .253. v .53). m .252 N .355. . .355. 035.5.» 9.9.0

 

N .59:— =o..ao._U gag—35v—

Am .038. 5.880 050.305.. 8.33. 5.323% 3.80:5 3.332 BEESéﬁN I c 2an

79

 

2588 pm:

toaqsm 32

250.50

Eta on:
26950

3.5 pm:

toaazm am:
2:250 pm:

toga—3m

Etna— n _ m
27,250 a _ I

toaazw am:

toaasm an:

2895 am:

toaazm «mm

276an mm:
toqum an:

2395 m _ :
toaasm 2:

ancEwﬁ \ guawoz

“$2.:ch \ 228m

anwzcwﬁ \ o>cwwoz

Eauwzcﬂm \ u>zawoz

EnocEwE \ u>sawoz
.52.:ch \ 328m

anEcwiﬁ \ oZEwoZ

EmocEwﬁ \ Q>Emom

~52:ch \ 03332

Eachwﬁ \ 323m

.caoc_=w_m5 \ ozawuz

.caoc_:w_m:_ \ 03332

ancEwE \ 03332
ascewa \ 3:58

E8E:w.m \ gumwoz

asceym \ 2,38“.

anEcwE \ o>_._mom

EmocEwE \ 03.5.9—

EsoEcw5 \ Q>Emcm

ESE—$5 \ 03:3;

:52:ch \ 03832
EnoEcwﬁ \ o>Emom

Echwﬁ \ 928m

anEcwE \ 358n—

.aaomﬁniawuc
-._.32m-§a25
1832255.
.._:§.£.§a§_a
war—azaaéiozn
b_.=a=O-.—Eon.b_muu>a

_a:o=au_=aw._O-.=e¢o.—m
.3_uo_c::uoh.56ao..m

«55.5%:
bugsOéEen—

 

 

ton—9%
m $852.5

tam—6
< $852.3

 

N .39:—

~ Ban-:—

8::

 

 

 

59556 $38M commwohwom I b 03.9

80

Results

Table 3 provides descriptive statistics and correlations for the study variables.
Tables 4 through 6 report the results of these analyses. Table 7 summarizes the results of
the regressions. To begin with, these results provide support for Hypothesis 1a. Table 4
provides support for the positive inﬂuence of experience depth on the rate of knowledge
creation, with Models 1 through 4 all indicating positive and significant relations for
depth variables. However, providing some boundaries on this relation, in the fully
saturated model (Model 9), the positive effects of depth on knowledge creation rate are
overwhelmed by other experience variables and fall insignificant. Moving on, the results
provide mixed support for Hypotheses 1b. Table 5 provides a fair amount of support for
the negative effect of experience depth on the creation of high impact knowledge, with
Models 3, 4, and 9 all showing signiﬁcant and negative relations between time-based
depth and the impact of knowledge creation. The pattern remains the same in Table 6 (2-
standard deviation impact), but the effects are weaker and fall insigniﬁcant. However,
demonstrating an interesting difference between time vs. quantity-based measures of
depth, the time—based measures of experience depth provided opposite results. In both
Tables 5 and 6, although the time-based measures supported Hlb, the quantity-based
measures were significant and positive (Models 1, 2, and 9), indicating that although
more time in the ﬁeld was negatively related to impact in knowledge creation, increasing
quantity of previous work was positively related to impact.

Moving to Hypothesis 2, the results provide mixed support for Hypothesis 2a.
Table 6 shows that one type of experiential breadth (organizational-breadth) is

signiﬁcantly negatively related to the rate of knowledge creation (Model 9). However,

81

contrary to Hypothesis 2a, technological-breadth is positively related to the rate of
knowledge creation (Models 5, 6, & 9 in Table 6). In terms of Hypothesis 2b, Tables 5
and 6 provide mixed support for the positive effect of experience breadth on the creation
of high impact knowledge. In this case, similar to the ﬁndings with experience depth,
again the two indicators of experience (in this case breadth) break apart with opposite
relations to knowledge creation. Technological-breadth in experience is positively related
to the impact of knowledge creation in Models 5, 6, and 9 in Table 5 and Models 5 and 6
in Table 6. However, organizational-breadth in experience is signiﬁcantly negatively
related to the creation of high-impact knowledge (Models 7 and 9 in Table 5 and Model 7
in Table 6). This is consistent with the correlation between this variable and these
dependent variables also being significantly negative (Table 3), the only core experience
variable so correlated.

Moving to Hypothesis 3, the results provide strong support for Hypothesis 3a.
Table 4 provides support for the positive inﬂuence of experiential diversity on the rate of
knowledge creation by firms in every model in which diversity is evaluated. Further,
there is also signiﬁcant evidence that these influences are subject to diminishing returns.
In Table 4, models 2, 4, 6, 8, and 9 all show a signiﬁcant positive relation between
different measures of experiential diversity and the rate of knowledge creation by firms.
Further, these effects are largely robust to the inclusion of measures of all types of
experience in the fully saturated model of experience (Model 9). Evidence of an inverted-

U relation7 is provided by the negative and signiﬁcant squared diversity terms in Models

 

7 Due to the small number of observations at high levels of diversity, additional tests of the signiﬁcance of
post-peak levels of the hypothesized inverted-U relationship were not possible. However. in addition to the
signiﬁcance of the squared diversity terms in the models. correlational analysis provides some additional
insight to support the found relationships. For example, I ﬁnd signiﬁcant positive correlations for diversity

82

 

2, 4, 6, 8, and 9. In addition, these analyses provide mixed support for Hypothesis 3b. In
Tables 5 and 6 there is some support for the positive influence of experiential diversity on
the creation of high impact knowledge in Models 6 and 9, where diversity in
technological breadth experience is positively related to both high-impact measures of
knowledge creation. In addition, in each of these models such benefits are subject to
diminishing returns with signiﬁcant negative squared terms. However, contrary to
Hypothesis 3b, Tables 5 and 6 also show several instances where diversity is significantly
negatively related to the creation of high-impact knowledge (Model 8 in Table 5 and
Models 2, 4, 8. and 9 in Table 6).
DISCUSSION

Overall, the two most signiﬁcant and consistent ﬁndings from these analyses are
that different types of organizational experience vary signiﬁcantly in their impact on
knowledge creation and that experiential diversity plays an important role in such
knowledge creation. In this sense, these results conﬁrm the need for more sophisticated
approaches to the conceptualization and measurement of organizational experience in

experiential learning research, as the influences of different types of experience can vary

 

in experience depth-quantity and experience depth-time with the rate of knowledge creation as .27
(quantity) and .05 (time) when all of the date is included. This supports the argument that experiential
diversity positively inﬂuences knowledge creation rate. However, at higher levels of diversity such positive
inﬂuences are argued to peak and become negative due to various problems created by such diversity.
Supporting this contention. the correlations for diversity observations that are above the mean in the sample
for experience depth-quantity and experience depth-time are signiﬁcant and negative (-.06 for quantity and
-. l4 for time). These become even more negative at higher levels, when we look at the few observed cases
of experience diversity that are more than a standard deviation above the mean level of diversity in the
sample. In these cases the correlations for experience depth-quantity and experience depth-time are -.28
for quantity and -.40 for time. Although inference from these findings is limited as this is only correlational
evidence, due to the small number of observations at these high levels of diversity. this further supports the
argument that diversity is valuable up to a point, but becomes an increasingly negative factor at high levels
of diversity.

83

widely (in terms of both significance and sign) in their effects on firm outcomes. The
implications of this research can be broken down into several observations.

First, this research has several implications for experiential learning theory and
research. To begin with, these findings support an overall contention of this paper that
depth and breadth of experience are distinct theoretical constructs that exhibit differential
predictive associations with key outcomes, supporting arguments of discriminant validity.
Further, in addition to demonstrating differences between depth and breadth, these
analyses also suggest that we may need to think more carefully about the convergent
validity of different measures of well-understood constructs, as these ﬁndings provide
evidence of the divergent inﬂuences of different types of experience depth (i.e. time and
quantity split inﬂuences). These split findings regarding quantity and time suggest a need
for more investigation into these two factors as potentially distinct constructs as well. In
addition, although experience depth is a well—studied construct (Argote, 1999), the
finding of a negative relation with the impact of knowledge creation represents a notable
exception to the primarily positive relations proposed between experience depth and firm
outcomes in the past (Yelle, 1979). Moving to experience breadth, these analyses
provided strong support for the inﬂuence of experience breadth on various dimensions of
knowledge creation, and again provide some interesting split ﬁndings between different
types of experience breadth. Although arguments can be made that the consistent
negative relations between organizational-breadth in experience and all three knowledge
creation outcomes could be the result of problems integrating old experience into new

environments, there is a need for more research into this phenomena to unravel its

84

mysteries. Together, these findings highlight how little we understand about experiential
breadth, and point to several new directions for additional research.

These results also highlight the important role of experiential diversity in
understanding the learning processes of organizations. This empirical evidence both
builds on and challenges the findings of Katila (2002) and Nerkar (2003) regarding the
mixing of old and new ideas in innovation. Specifically, although these results conﬁrm
the positive inﬂuence of diversity on the rate of knowledge creation (Table 4), they also
challenge the idea that integrating old and new ideas is a positive opportunity for
organizations by showing negative inﬂuences of such integration on knowledge creation
impact (Table 4 in particular). Further, in addition to the investigation of ‘age of
knowledge’ being integrated, the current study also found support for the positive
inﬂuence of diversity in breadth of experience, a theoretical path not previously explored.
Together, these findings suggest that although experiential diversity has promising
potential as a construct in experiential learning, these mixed effects make it clear that we
still have much to learn about diversity and how it is managed.

In addition, this research also provides some of the ﬁrst empirical support for core
rigidities arguments (Leonard-Barton, 1992). Although core rigidities arguments are well
established in the strategy literature, there has not been very much empirical research into
this phenomenon. By assessing core rigidity ideas directly through the analysis of actual
experience (Hlb), these results bolster the theoretical arguments for core rigidity’s
negative inﬂuences in organizations, though they also demonstrate some of the

limitations of such ideas (only partial support).

85

Finally, also notable for the experiential learning literature is the sensitivity of
these results to different types of experiential measures (time-based vs. quantity-based as
well as technological vs. organizational breadth). Although previous research has
employed each of these methods, few studies examine all of these different
operationalizations of experience together. By examining the independent and joint
inﬂuences of different operationalizations, it became clear that how experience is
conceptualized and measured has an important effect on the results. This may serve as a
warning to future research.

In addition to the implications of this research for experiential learning research,
this research helps to answer the call from Nonaka and colleagues (Nonaka, et al. 2006)
for new research into the origins of knowledge creation. This study helps to bridge the
gap between experiential learning research and knowledge creation research to specify
the nature and complexity of the experiential antecedents to knowledge creation in
organizations. In doing so, these results suggest that simpliﬁed and implicit approaches to
experiential antecedents are probably not sufﬁcient to understand complex dynamics of
knowledge creation, given the significant divergence in effects found here. Further, these
analyses highlight the importance of considering different dimensions of knowledge
creation outcomes (rate and impact), as different experiential origins were found to have
divergent inﬂuences on different dimensions of knowledge creation outcomes.

This research also has several implications for practicing managers and
organizations. First, experience can be a double—edged sword. Although the accumulation
of experience can provide considerable beneﬁts to organizations (such as the positive

relation found between experience depth and knowledge creation rate), the same

86

experiences can also have notable negative consequences as well (core rigidities and less
knowledge impact). Additionally, these results suggest that executives should think about
experience in terms of more than just depth of experience, particularly given the ﬁndings
regarding breadth of experience and the creation of high—impact knowledge, and that
experiential diversity can be a significantly positive force in organizations. Although
difﬁcult to predict practical business realities often create experiential diversity (turnover,
industry dynamism, etc.), such diversity should be fostered and can lead to positive
organizational outcomes.

Finally, this research also highlights several future directions for additional
research. Although this study primarily examined two different types of experience, there
is a considerable opportunity to investigate the inﬂuence of additional dimensions of
experience (challenge, timing, context, density, etc.) on knowledge creation. Building on
this, research could also begin to explore the differential effects of different types of
organizational experience (at multiple levels) on other performance and strategic
outcomes (competitive dynamics, different types of ﬁnancial performance, alliance
formation and value creation, survival, mergers and acquisitions, etc.). Another new
research direction would be to look at these issues in the context of mergers and
acquisitions. For example, it is possible that the negative ﬁnding regarding organizational
breadth of experience in this study could include two different types of breadth, forced /
involuntary organizational breadth as the result of mergers and acquisitions, and
voluntary breadth as a result of proactive scientist mobility. Differentiating these two

different types of experience breadth, and the different challenges associated with these

87

moves (culture shock, scientist motivations, financial rewards, etc.), could be fruitful
ground for new research.

In addition, although this research explored a large number of technologies in a
single industry, it is possible that different types of experience could be differentially
important at unique points in the lifecycle of a research domain. For example, it might be
possible that breadth of experience is the primary driver of success at early stages of a
technological lifecycle, but at later points experiential depth becomes key to marching the
technology forward. Additionally, future research could also break knowledge creation
down further based on the actual novelty of the innovation. For example, although this
study examined high-impact patents as a speciﬁc outcome, it would also be possible to
explicitly study the experiential antecedents of distinctively incremental knowledge
creation or only knowledge creation that is a continuation of a firm’s own previous work.
Finally, future studies could also consider the inﬂuence of different types of experience
on intermediate knowledge creation processes (conﬂict, discourse, knowledge sharing,
search processes, use of technology), steps that were beyond the scope of this research
but could provide more insight into the micro processes of experience and knowledge
creation.

In conclusion, this study aimed to take new steps toward a deeper understanding
of the inﬂuences of different types of experience on knowledge creation in organizations.
By investigating the inﬂuence of research scientist experiences in the pharmaceutical
industry, some questions were answered and new questions were raised. Hopefully this
study will encourage more research into some of these complex. important, and

interesting questions.

88

CHAPTER 3
CREATING BLOCKBUSTERS: UNIVERSAL AND CONTINGENCY

PERSPECTIVES ON EXPERIENTIAL LEARNING AND PERFORMANCE IN
HOLLYWOOD

OVERVIEW

In this study the relation between different types of organizational experience and
different types of performance is explored. Building on experiential learning theory, two
perspectives are advanced. First, a universalistic approach is developed from the
traditional learning literature, building on the logic that all experience is good experience.
Second, a more contingency-based approach is hypothesized, in which the inﬂuence of
different types of experience on performance can vary, and even be negative in some
cases. Data from the Hollywood motion picture industry is used to evaluate these
perspectives. Results provide only limited support for a universalistic perspective,
suggesting that the universalistic approach employed in some previous experiential
learning research may be only one piece of the puzzle, and could lead to misleading
answers to important questions. Instead, the current research provides strong evidence for
the importance of various contingencies in this relation, including the type of experience,
the type of performance assessment, and the locus of performance evaluation (internal

reference points or market-based comparisons).

89

Although experiential learning has been studied in organizations for many years,
until recently much of this research has taken a simplified view of organizational
experience as consisting of repeated exposure to similar situations over time (Yelle,
1979). In recent years this research has branched out to some degree, to better consider
how organizations might also learn from doing ‘something else’ (Schilling, Vidal,
Ployhart, & Marangoni, 2003), the implications of more local and distant learning
(Rosenkopf & Nerkar, 2003; Nerkar & Roberts, 2004; Ahuja & Lampert, 2001), and the
inﬂuences of specific types of experience (Haunschild & Sullivan, 2002; Hoang &
Rothaermel, 2005; Haleblian & Finkelstein, 1999; 20110 & Singh, 2004). This progress
has marked a significant departure from the more narrowly focused traditional leaming-
by-doing literature (Yelle, 1979; Argote, 1999) and has highlighted how little we
understand, both theoretically and empirically, about how organizations learn from
different types of experience.

In the current study this departure is further advanced by exploring a basic
theoretical question in organizational learning. Speciﬁcally, I examine the question of
whether different types of organizational experience have a universal or contingent effect
on different types organizational performance. A universal perspective is the simplest and
most intuitive explanation, building on the traditional logic that any experience is
valuable experience. In contrast. a more contingent perspective would suggest that the
value of different experiences vary and that in some cases certain types of experience
might have little or even negative inﬂuences on some types of performance.

To evaluate these two potential perspectives on the relation between experience

and performance, I will examine the inﬂuences of two different types of experience in the

90

production of major motion pictures in Hollywood. On one hand, I will argue from a
universalistic perspective to suggest that any accumulated experience, regardless of type,
will have a positive inﬂuence on film performance. On the other hand, I will also offer
hypotheses from a contingency perspective to suggest that experience depth (repeated
exposure to similar situations over time) and experience breadth (exposure to a variety of
different situations over time), two different types of experience, will have divergent
inﬂuences on the creation of blockbuster performance. In this contingency logic, I build
from ideas from creativity research to argue that experience breadth will have a
signiﬁcant positive inﬂuence on the creation of blockbuster movies. Then, building from
core rigidities research, I will argue that experience depth can negatively inﬂuence the
probability of creating a blockbuster movie. In this sense, these arguments would suggest
that these two different types of organizational experience have opposite inﬂuences on
blockbuster performance.

Further, as organizational performance can be theoretically and empirically
conceptualized in many ways (Dess & Robinson, 1984; March & Sutton, 1997),
particularly in this context (Ferriani, Cattani, Baden-Fuller, 2007), several perspectives
on performance will be considered. First, I begin with the most traditional evaluation of
performance in this industry, an overall performance measure of commercial success —
box office revenue created by the US. release of a ﬁlm (a continuous variable). Second, I
build hypotheses to explore the divergent inﬂuences of different types of experience on
the creation of blockbuster hits as an indicator of performance (a binary variable), an
important measure of success in this industry given the large percentage of box ofﬁce

revenue generated by the most successful ﬁlms (as shown in the Lorenz curve for this

91

industry in Figure 2)8. Third, although firms compete to survive in a competitive
marketplace, performance can be evaluated as success relative to the market (external
comparisons) or relative to a ﬁrm’s own previous performance (behavioral theory’s
internal reference points). As such, both of these perspectives will be explored. Finally, in
the Hollywood context success can be conceptualized in both a commercial and creative
sense, although the two are not always well correlated. Thus, to examine performance in
this industry, both subjective evaluations of creative success (critical ratings) and
objective financial performance (box ofﬁce revenue) will be considered.

In advancing these different ideas, I aim to make several contributions to the
literatures on organizational learning and performance. First, by outlining the differences
between universal and contingency perspectives on the link between experiential learning
and performance I help move forward our theoretical understanding of how
organizational experience can inﬂuence performance. Second, by hypothesizing from
both universal and contingency perspectives rather than advocating one over the other I
can investigate empirically both perspectives on this link between organizational
experience and performance. Third, by including multiple types of experience (depth and
breadth) this research continues the push to expand beyond narrower approaches to
organizational learning, to consider both the depth and breadth of organizational
experience independently. Finally, by exploring a variety of different theoretical
conceptualizations of performance, I hope to open new doors to the understanding of the
link between organizational experience and different forms of performance.

With the goal of exploring these issues, this study is structured as follows. I begin

by developing the theoretical differences between universal and contingency perspectives

 

8 See a more complete description of the importance of blockbuster hits in this industry on pages 99- l0l.

92

to link organizational experience and organizational performance. I then proceed to
develop hypotheses from each of the perspectives, first arguing that any experience helps
to improve performance, but then moving on to hypothesize that the type of accumulated
experience might matter when seeking to explain blockbuster performance. These ideas
are then tested empirically in the Hollywood motion picture industry, providing a rich
data context for the evaluation of the proposed hypotheses. The paper concludes with a
discussion of the implications of this study for organizational learning theory and also for
practice, in addition to identifying potential directions for future research.
THEORY AND HYPOTHESES

Universal and Contingency Perspectives on Experiential Learning and Performance

Exploring how experiential learning inﬂuences organizational performance is
fundamental to the study of organizational learning because it helps to motivate research
into organizational learning as a field of research. Reﬂecting its fundamental nature, this
link has been well explored in research (largely in economics) that has clearly established
the link between organizational experience and economic outcomes such as learning rate
and production efficiency (Argote, 1999; Dutton, Thomas, & Butler, 1984). However,
although the literature has branched out in a variety of ways to explore new contexts in
which such learning occurs (from nuclear power generation to pizza-making [Darr,
Argote, & Epple, 1995; Zimmerman, 1982]), new methods of study (from ﬁeld to lab and
simulation [Muth, 1986; Schilling, et. al, 2003]), and new outcomes (from learning rate to
the reduction of errors and ﬁrm survival [Baum & Ingram, 1998; Haunschild & Sullivan,
2002; Hirsch, 1952]), there has been little systematic evaluation of the nature of the link

between different types of experience and overall organizational performance.

93

Reasonable arguments can be made to support either a universal or a contingent
perspective to link experiential learning and organizational performance. From early
research (e.g. Alchian, 1963; Wright, 1936) through recent work (e.g. Darr, et. al, 1995),
the universal perspective has been the predominant approach to this relation, and
increasing experience was argued to move a ﬁrm along the learning curve. Examples
such as Preston and Keachie‘s (1964) work on decreasing labor costs in the production of
radar equipment and Darr, Argote, and Epple’s (1995) work on decreasing production
costs in pizza making are typical of the universal approach. In these examples, and many
others, the relation between organizational experience and organizational performance is
simpliﬁed and represented as a direct positive inﬂuence that is subject to diminishing
returns.

Nonetheless, additional reasonable arguments can also be made to support a
contingency perspective of the organizational experience to performance link. For
example. research on learning traps and biases (e. g. propinquity traps, familiarity traps,
maturity traps, and superstitious learning) suggest that in some cases experience can
negatively inﬂuence organizational performance (e.g. Ahuja & Lampert, 2001; Levitt &
March, 1988). Further, empirical examples such as Baum and Ingram’s (1998) work on
organizational survival in the hotel industry and Nerkar and Robert’s (2004) work on
innovation in the pharmaceutical industry are also typical of a more contingent approach.
In the Baum and Ingram (1998) research the contingency of interest was the local or non-
local nature of the experience accessed by hotels. In Nerkar and Roberts’ (2004) research
the contingency was the type of product or market experience accumulated. In each of

these cases. it was argued that the relation between organizational experience and

94

organizational performance is contingent upon some characteristic of the experience or
situation. In the case of learning traps, some organizational experiences are argued to
specifically impede the achievement of firm goals by predisposing ﬁrms to sub-optimal
behavior. In this way, the theoretical argument implicitly being advanced is that although
experience may be good, it could come with conditions attached.

Rather than forcefully advocating one perspective over the other, I will argue that
both universal and contingent perspectives could be legitimate approaches to the study of
organizational learning. In the following sections each of these perspectives will be used
to motivate hypotheses. In this way, both universal and contingent perspectives are
advanced as potentially useful and informative, allowing the data to provide further

insight into this issue.

All Experience is Good Experience? The Universal Benefits of Experience for
Performance

The idea that all experience is good experience is a common reﬂection that has
become well established in the literature. This perspective takes an undifferentiated view
of experience as a positive opportunity for learning, building on Kolb’s (1984) theoretical
framework applied to the organizational level of analysis (Popper & Lipshitz, 2000). The
theoretical argument for this perspective, developed by Kolb, is that any and all
experience provides the raw material and impetus for observation, reﬂection, forming
abstract concepts, and testing out new ideas to improve future interactions (Kolb, 1984).
Synthesizing these ideas and bringing them formally to the organizational level of
analysis, Popper and Lipshitz (1998; 2000) have argued that such experiential learning

processes also work in organizations. The idea is that as organizations gain experience,

95

these experiences give organizations the opportunity to improve their performance over
time by learning to avoid mistakes from past failures (Vaughan, 1999), learning new best
practices (Schwandt & Marquardt, 2000), learning how to manage their alliances
(Simonin, 1997), learning to improvise effectively (Barrett, 1998), learning to better
identify technological opportunities (Dodgson, 1991), learning to reduce errors (Argote &
Darr, 2000), and learning from a host of other experience- generated insights to improve
performance (Bapuji & Crossan, 2004). Such improvements could be in terms of
efficiency—related performance (Argote & Epple, 1990), error-related performance
(Argote, 1999), quality—related performance (Argote, I993), survival (Baum & Ingram,
1998) or any other type of ﬁrm performance outcome (Dodgson, 1993). The deﬁning
theoretical argument is simply that increased experience improves performance.

As noted earlier, a consistent stream of empirical research has been presented
through the years to support this undifferentiated perspective, with much of the focus
being directed at the nature of the diminishing returns for such learning (the learning
curve) in different settings, and factors that inﬂuence these diminishing returns (for a
review see Argote, 1999). Nonetheless, the primary argument in this research has been
that learning from experience does occur and that increasing experience promotes
organizational performance.

In this tradition, the most common representation of organizational experience has
been the accumulation of similar experiences over time, usually represented by
cumulative output (experience depth). Building on this, recent research has continued the
universal approach by suggesting that not only do similar experiences accumulate to

improve performance, but organizations may also improve performance by learning from

96

doing ‘something else’ as well (experience breadth). Although this was a lab-based
experiment (Schilling, et al., 2003), it represents a yet more pronounced endorsement of
the universal perspective, suggesting that any experience, even if it is experience that is
not a direct replication of previous experiences, can improve performance.

In this research I build on this universalistic idea, implicitly advanced by
organizational learning scholars through the years, that organizations learn from all of
their experiences to improve overall performance. Thus, the argument is that regardless
of the type of experience gained, whether it is depth of experience resulting from the
accumulation of similar experiences or breadth of experience resulting from the diversity
of such experiences, increasing levels of experience will universally improve
organizational performance.

HI: All forms of organizational experience in the industry will promote improved

performance.

A Contingent Perspective on Experience: Tradeoffs for Blockbuster Performance

In contrast to the previous section, a more contingent perspective would suggest
that not all experience will improve performance, and that various contingencies could
interfere in this direct link. As outlined earlier, this contingency approach has been
reﬂected in several recent research findings, from Baum and Ingram’s (1998) work on the
local versus non—local locus of experience and the timing of such experience as
contingencies, to Nerkar and Robert’s (2004) suggestion that the type of experience
(product versus market) can act as a contingency to amplify or weaken the effects of
experience on performance. Although neither Baum and Ingram (1998) nor Nerkar and

Roberts (2004) explicitly characterized their research as taking a contingency

97

perspective, both studies took a contingent approach to experience by describing the
relation between experience and performance as subject to different conditions. In the
case of Baum and Ingram (1998), the timing and locus of experience were found to be
important in the survival of Manhattan hotels. In the case of Nerkar and Roberts (2004),
complementarities between technological and product-market experience were theorized
to create a joint positive synergy and improve the success of new product introductions,
contingent on the presence of both types of experience together in the same ﬁrm.

In this study I build on these contingency approaches to theory on the link
between organizational experience and performance, but take the contingency arguments
in several new directions. First, I focus on new experience types (the differences between
experience ‘depth’ and ‘breadth’) to suggest that these two different types of
organizational experience also act as a contingency in the link between experience and
performance. Further, instead of arguing for an ampliﬁcation or weakening effect, I
suggest that these two different types of experience can have opposite effects. In addition,
I extend the contingency to the performance outcome as well to explore the dependent
variable as a potential contingency in addition to the independent variables. The
performance outcome I focus on is the achievement of blockbuster performance rather
than just overall performance, new product introductions, economic efﬁciency, or
survival. As will be described below, I chose blockbuster performance specifically
because this organizational performance outcome has the opportunity to provide an
interesting theoretical test for experiential leaming’s effects, as research from two
different research traditions motivate support for two different directions in anticipated

performance effects.

98

% of Revenue

Figure 2 — Lorenz Curve for Box Office Revenue in Hollywood

 

100

80 a

60

4o

20 ~

 

 

 

 

 

 

o 20 4o 60 80 100
°/. of Films

Note: Based on data from IMDB.com for the years 1993-2000.

Although ‘blockbuster performance’ is not a traditional dependent variable in the
organizational learning literature, in the empirical setting of this research blockbuster
performance (i.e. performance significantly above typical performance in the industry)
plays an important role in the survival and thriving of movie producing organizations.
Indeed, some have argued that blockbuster success is critical to the viability and success
of many Hollywood production companies as one big movie success is often enough to
set up the ﬁrm for years of future work (Stringer, 2003). A Lorenz curve (shown in
Figure 2) demonstrates how the distribution of box ofﬁce revenue between ﬁlms in a year
makes blockbuster movies so important and attractive. In fact, this increased focus on big
successes is not unique to the movie production business, as similar arguments have been

made in other high fixed-cost industries, such as the pharmaceutical industry where the

99

big profits of blockbuster drugs provide the capital for an organization to engage in a
wide range of less lucrative therapeutic endeavors. In other industries similar dynamics
occur, such as book publishing (e. g. creating the ‘Da Vinci Code’ brand), consumer
goods (e. g. beer and home products), and electronics (e. g. the iPod phenomenon), where
a blockbuster success can create a new brand that generates a disproportionate positive
impact for the ﬁrm that creates it relative to typical ﬁlm performance. In the context of
Hollywood, blockbuster movies provide several important beneﬁts to the firms that
produce them, both in terms of financial rewards and status enhancement. For example,
out-of—theatre commercialization opportunities (such as toy sales and licensing
arrangements) accrue disproportionately to blockbuster movies, providing enhanced
recognition and revenue generation for the firms that produce such movies. Further, the
opportunity sets available to such ﬁrms may become enhanced, as these ﬁrms may get
preferential access to talent, scripts, partnership arrangements, and studio distribution
deals which can have long-term implications for the viability and survival of such firms.
As a result, blockbuster performance can be viewed as an important organizational
performance outcome for many industries, and one that has not been well explored in the
literature.
Experience Breadth and Blockbuster Performance

To begin with, increasing levels of experience breadth is argued to promote
blockbuster performance. This is hypothesized for several reasons. First, research on
creativity suggests that diverse experience can provide an organization with exposure to
unique sources of knowledge which can then be infused into future projects to increase

the potential for big ideas. Speciﬁcally, such research has found that breakthrough

100

performance is most likely to result from the integration of previously unconnected ideas
in different domains (Kelly & Caplan, 1997). In an ethnography describing the work of
IDEO, a product design company, Hargadom and Sutton (1996) described how the
location of this ﬁrm as a central technology broker allowed the firm to identify high-
potential opportunities not clearly visible to other ﬁrms. Similarly, diverse organizational
experiences also are likely to provide firms with access to opportunities with higher
uncertainty but higher return potential, as an artifact of the boundary spanning role of
such firms, increasing the potential for creation of high-potential projects. Second,
Nemeth (1986) suggests that the exposure to qualitatively different technologies over
time has the effect of expanding an inventor’s cognitive models and can help them to
make novel connections that they may not have otherwise recognized. As creative
endeavors are often inherently uncertain processes that draw heavily on an individual’s
developed tacit knowledge (Polanyi, 1967) and intuition (Policastro, 1995), expanded
cognitive models drawn from diverse experience are likely to support such creative work.
Finally, access to diverse knowledge from multiple domains combined with such diverse
cognitive model exposure helps to aid such individuals in knowledge recombination
efforts (Fleming, 2001), raising the potential for blockbuster synergistic combinations of
related but distinct ideas. Together, these ideas support the overall contention that
increasing breadth in an organization’s experience set should better equip it to achieve
blockbuster performance.

H2: Increasing breadth in organizational experience will lead to an increasing

likelihood of blockbuster perfonnance.

Experience Depth and Blockbuster Perfomzance

101

In addition to the positive inﬂuence of experience breadth, we can also consider
the effect of experience depth on the creation of blockbuster performance in Hollywood.
In contrast to the previous hypothesis, I posit that experience depth has a negative effect
on the production of blockbusters, and that with increasing levels of depth in
organizational experience blockbusters become less and less likely. Thus, the effect of
experience on the creation of blockbusters is proposed to be contingent on the type of
experience accumulated such that experience breadth promotes blockbusters but
experience depth reduces the chance of achieving a blockbuster.

This hypothesis is based on several related ideas. First, increasing levels of depth
in prior organizational experience has been argued to create core rigidities in the way
organizations approach future problems (Leonard-Barton, 1992; Dougherty & Heller,
1994). Speciﬁcally, with increasing levels of historical dominance of one speciﬁc
knowledge domain inside a ﬁrm, the ideas and processes involved in this recurring area
are elevated to a high status position in the ﬁrm, and outside experiences receive less
attention and representation in ﬁrm strategic decision making (Leonard-Barton, 1992). As
a result, over time with increasing levels of specialization in a domain, the values of the
firm become aligned in the pursuit of this work to the exclusion of other ideas, and the
dominant logic of the organization’s senior management can become entrenched
(Hambrick, Geletkanycz, & Fredrickson, 1993; Prahalad & Bettis, 1985). The result over
time is a rigidity that is likely to constrain the integration of creative new ideas that would
help to create blockbuster performance. Second, increasing levels of depth in a particular
knowledge domain can have the unintended consequence of developing an inside-the-box

mentality, or the ‘not-invented-here’ syndrome (Katz & Allen, 1982), and can grow into a

102

tendency to look at things according to the conventions of one's own profession, not
learning broader points of view. As such closed-off thinking is contrary to the factors that
positively inﬂuence creativity (Amabile, Conti, Coon, Lazenby, & Herron, 1996), such
deep but narrow mentalities are likely to predispose the firm to a variety of learning traps
(Ahuja & Lampert, 2001) and reduce the likelihood of such ﬁrms achieving breakthrough
innovation. Finally, with increasing depth of experience over time ﬁrms develop norms
and routines to support such typical operations. Norms have been called one of the least
visible but most powerful forms of social control over human action (Bettenhausen &
Mumighan, 1985), and with increasing levels of organizational experience in a product
domain, norms grow to both support and eventually socially require standardized work
practices. As norms and routines create inertia (Harman & Freeman, 1984), over time
such norms act to constrain the organization toward more of the same and make it
difﬁcult to create the novel projects that lead to blockbuster performance. Together, these
arguments support the following hypothesis.

H3: Increasing depth in organizational experience will lead to a decreasing
likelihood of blockbuster performance.

METHODS
Data and Setting
The US. motion picture industry was selected as the setting to study the proposed
hypotheses, providing an information-rich setting to evaluate the effects of different types
of experience on organizational learning and performance. This is an industrial context
where organizational learning is prized, and firms are eager to transfer their learning from

one project to future endeavors through both formal and informal mechanisms. as the

103

costs of failure are significant but the rewards for success are high. To explore
experiential learning in this context, data on every movie commercially released in the
United States during the years from 1993-2000 were collected to create organizational
experience variables for the production companies involved in creating these ﬁlms.
However, due to signiﬁcant differences in the scope and mission of different
production companies during this time, it is difficult to compare the learning that takes
place in different types of production companies. For example, although production
companies are responsible for the creation of new ﬁlms, in some cases production
companies are formed as more of a ﬁnancial backer than a creative production partner
and in other cases they may simply represent the interests of a foreign government or
agency involved with the film. Thus, to help to ensure consistency in the nature and
scope of production companies included in the study, I restricted the sample for primary
analysis to only include movies made by production companies that were afﬁliated with
and distributed by a major Hollywood studio. These studio—afﬁliated production
companies represent the primary developers of mainstream movies in Hollywood and
compete with each other to create successful movies for the public. As the goal of this
research is to understand how different forms of organizational experience inﬂuence
performance, these firms represent the primary sources of such learning in Hollywood.
By restricting the sample to include only these well deﬁned production companies, I help
to ensure that the experiential inﬂuences being measured represent the experiences
possessed by the primary technical and creative forces behind the production of each new
movie in the dataset. In total, the data included 2,888 ﬁlms released between 1993 and

2000. Of these films 1,149 unique movies were released by major Hollywood studios,

104

produced by 269 unique production companies. As the focus of this research is on the
accumulation of experience, the analyses presented in the tables report only the 559 films
released during the five years 1996-2000, the years in which at least three years of prior
data existed to assess prior experience.9 The data from 1993 to 1995 and the ﬁlms that
were not released by major Hollywood studios were only used for experience
accumulation calculations and robustness checks.

Independent Variables

The independent variables fall into three categories, following the delineation
between overall experience, experience depth, and experience breadth developed in the
theory and hypotheses section of the paper. All measures were calculated at the
organizational level of analysis and updated with each new movie that an organization
produced.

Overall Organizational Experience. The first independent variable measured the
overall level of movie-making experience for each ﬁrm prior to working on a new movie
project. The overall level of production company experience was measured by counting
the number of ﬁlms a production company had created in the three years prior to working
on a new project. This measure was undifferentiated, in that it did not take into account
the type of movie being made (adventure, horror, etc.) or any other characteristic of the
experience, and instead was just a generic overall measure of experience in the movie

business. This was the primary measure for the analysis of the idea that any experience is

 

9 This 3-year time frame was used to focus analyses on the experiences most relevant to each ﬁrm. and
taking into account that knowledge gained from older experiences decays (especially in the presence of
new experience), particularly for project-based organizations with signiﬁcant turnover. For robustness.
alternative models were also speciﬁed to include experiences as old as 5 years to assess the differences in
results between models that focused on more proximal and more distal experiences. There were no
systematic differences in results. although the 3-year results are a little stronger.

105

valuable experience argument. The average firm in our sample had produced over eight
movies in the three years prior to producing a new movie.

Depth of Organizational Experience. The second independent variable measured
the depth of organizational experience for a firm prior to working on a new movie
project. Depth in a production company’s experience was measured by counting the
number of previous ﬁlms a production company had created in a particular genre in the
previous three years. To signal to movie audiences what a movie is about, upon release
every movie is categorized into a genre, and in some cases a secondary or tertiary genre,
to specify the category of movies to which it belongs. These different categories reﬂect
unique knowledge domains, as the skills needed to make a ‘comedy’ differ signiﬁcantly
from those needed to make an ‘adventure’ movie. Thus, this measure simply provides a
count of the number of movies a production company has made in a particular genre. In
the case of movies that fall into more than one genre, the production company’s
experience making ﬁlms in any of the genres associated with the ﬁlm are counted. Thus,
if the focal film is an adventure-comedy, the depth of a ﬁrm’s experience would be
assessed as the number of ﬁlms the ﬁrm had previously produced that were either
adventure films or comedy films”). The 15 primary movie genres are provided in
Appendix B, and are based on the genres identiﬁed by Miller and Shamsie (1996). The

average firm in our sample had produced about 2.5 movies in the speciﬁc genres

 

'0 An alternative approach to the measurement of experience depth would be to only count depth as
experience making the EXACT same type of ﬁlm over time. In other words, if the current ﬁlm being
produced was a crime-comedy. only previous crime-comedies would be counted as relevant previous
experience. and prior experiences making just ’comedies’ or ‘crime’ ﬁlms would not be counted. Models
analyzed using these exact-matched experiences followed the same pattern of results as those described
here. but as this exact-matched measure was less inclusive of relevant previous experiences in the same
genres. the experience depth measures reported were those that measured all previous experience in the
focal genres.

106

represented in a new project over the three years prior to producing a new movie, and
ranged in experience depth from 0 to 17 previous films.ll

Breadth of Organizational Experience. The third independent variable measured
the breadth of an organization’s prior experience. Breadth in a production company’s
experience was measured with a count of the number of unique genres in which a
production company had produced movies in the previous three years. These different
categories reﬂect potentially unique knowledge domains, and represent the breadth of an
organization’s experience with different types of movie products. Again, in the case of
movies that fell into more than one genre, the production company was credited with
experience for all of the genres associated with a particular ﬁlm. Thus, if a ﬁlm was
classified as a crime-drama, a firm would get credit for both ‘crime’ and ‘drama.’ Again,
the 15 primary movie genres are provided in Appendix B, and are based on the genres
provided by Miller and Shamsie (1996). The average ﬁrm in our sample had produced
movies in about ﬁve different genres over the three years prior to producing a new movie.

and ranged in experience breadth from 0 to 12 unique genres.

Dependent Variables
A significant advantage of choosing the Hollywood motion picture industry as a
setting for this research was the availability of unique and consistent performance criteria

for evaluation over time. As organizational performance is a complex construct with

 

” Similar to the issues in the previous footnote. another potential question regards the number of ﬁlms that
were classiﬁed as representing more than one genre (74% of the films). In calculating the experience of
production companies. these firm’s experience making prior ﬁlms in any of the genres of the current movie
were added together to give a full picture of the accumulated experiences of such ﬁrms. However, this
could potentially inﬂate the experience numbers of firms that had produced many ﬁlms that were classiﬁed
with multiple genres. To assess the influence of this issue. alternative measures were calculated that
averaged the experiences of a ﬁrm making films in each genre represented in a current film rather than
adding them together. The results were identical, in many cases identical to the third decimal place.

107

many different indicators (March & Sutton, 1997), several different indicators of
organizational performance were analyzed to provide a rich view of the effects of
organizational experience on such performance. These various indicators of
organizational performance fall into two categories — overall performance and
blockbuster performance.

Overall Performance. The ﬁrst category of dependent variables was a set of
indicators that assessed overall performance. In Hollywood, overall performance is a
complex concept. For most, success is measured by the money generated by a ﬁlm - its
commercial / popular success at the box ofﬁce. For others, however, success for a film is
measured by the critical acclaim generated by a ﬁlm, the degree to which the ﬁlm is
viewed to be a masterful artistic creation. To assess both of these types of success, overall
performance for ﬁlms was assessed in both ways to measure both the commercial and
critical success of a ﬁlm. First, overall commercial success was assessed by examining
the total box ofﬁce revenue generated (in US. dollars) from the US. release of each of
the firm’s ﬁlms. This is the most widely used criterion for performance in the motion
picture industry. Second, critical movie success across movie projects was assessed using
the evaluations of professional movie critics from major news media outlets in the United
States. These evaluation ratings of professional critics were gathered from an industry
source of such ratings, rottentomatoes.com, and averaged to create a composite critical
rating score. In these data, movies were evaluated by an average of 41 professional

. . o . . ‘
critlcs'“. As an example, a representative movre from the dataset was ‘Enemy of the

 

'3 An alternative approach to the use of all professional critics reported by rottentomatoes.com was the use
of a selective group of the top. most nationally inﬂuential movie critics from around the country (the cream
of the crop of professional movie critics). Analyses that utilized only the ratings of this more selective

108

State,” an Adventure-Thriller starring Will Smith and Gene Hackman, which was
evaluated by 65 professional critics ranging from Peter Travers of Rolling Stone
magazine to James Sanford of the Kalamazoo Gazette.

Blockbuster Performance. The second category of dependent variables included
indicators that assessed blockbuster performance. In this case, again both critical and
commercial success metrics were used to provide alternative views of stand-out
performance. A ‘critical blockbuster’ was identiﬁed as a film with a cumulative
professional critic rating that was at least one standard deviation above the mean critical
rating for films in a particular genre in a given year. This was dummy coded as a ‘1’ for
critical blockbusters and ‘0’ for other ﬁlms. The second measure - a ‘box office
blockbuster’ — was assessed similarly, but using box ofﬁce revenue instead of critical
ratings. This was dummy coded as a ‘1’ for blockbuster hits and ‘0’ for other films”. In
addition, although these two measures assess the degree to which a focal project resulted
in blockbuster performance relative to other films in the market in a given year (overall
market blockbuster), the behavioral theory of the firm (Cyert & March, 1963) argues that
it is also important to consider the degree to which a ﬁlm was a blockbuster for the
production company that produced it (ﬁrm-specific relative blockbuster), regardless of
how it compared to other films in the overall marketplace. Thus, two additional variants
of blockbuster performance assessed the degree to which a particular film was a ‘critical

blockbuster’ or ‘box ofﬁce blockbuster’ relative to the performance of the production

 

group of critics provided consistent results with those reported in the results tables. but with a smaller
sample of films. as these top critics evaluate a smaller percentage of Hollywood ﬁlms.

‘3 To assess the robustness of these genre-specific blockbuster measures to the comparisons of ﬁlms to all
ﬁlms in the marketplace. rather than just in the same genre. alternative models were examined that
compared a ﬁlm's performance to all other ﬁlms in the marketplace (absolute blockbusters). These were no
systematic differences.

109

company’s own previous work. Firm-relative blockbuster performance was assessed as
box ofﬁce revenues or critical ratings more than one standard deviation above the ﬁrm’s
previous performance”.
Controls

To control for potential alternative explanations of results or the effects of other
inﬂuential factors, several additional variables were included in the analyses as control
variables. First, the budget of a ﬁlm was included as a control, as the money put into
making a ﬁlm has an important role in the success of the film, particularly for the
commercial success measures (i.e. big budget versus small budget ﬁlms). Second,
dummy variables for the year a movie was produced were included to control for
differences over time in models of overall performance (Table 9). Time was already
controlled in the blockbuster-based dependent variables by assessing blockbuster status
within year, so period controls were not included in Tables 10 or 11 (though the inclusion
of such controls had no effect on the results). Third, the number of screens the ﬁlm was
initially released on to audiences was added to control for the impact of breadth of release
on performance. Fourth, in cases where the core production companies (with studio
afﬁliations) brought on other production companies as partners to create a new film, a
variable measuring the number of partners involved was included to control for the

inﬂuence of such partners.

 

'4 A ﬁrm's previous performance was judged by the performance of their 5 most recent ﬁlms. If the ﬁrm
had not produced ﬁve previous films. the comparison would be the performance of the current ﬁlm relative
to all previous films the ﬁrm had produced. Alternative versions of this measure compared the performance
of the current ﬁlm to the previous 3 films the firm had produced. or all of the ﬁlms a firm had produced in
the previous 3 to 5 years. Further. as calculating a standard deviation on only a few films could be
comparing current performance to a very small number of previous films ( e. g. the standard deviation of
performance on only 2 or 3 previous films), alternative models were also evaluated that only analyzed firms
that had 5 or more films from which to calculate a standard deviation. All comparison windows provided
identical results, with only small changes to the actual effect coefﬁcients.

110

Analyses

The data were analyzed using panel regression in the statistical package Stata. For
the regressions involving blockbuster performance, the binary nature of the dependent
variable required the use of panel-based logistic regression (Greene, 2003). Hausman
tests (Hausman, 1978) were used to determine if firm-speciﬁc ﬁxed or random effects
estimators were appropriate for the various models by testing the null hypothesis that the
coefﬁcients estimated by the more efficient random effects estimator were consistent
with those estimated by the ﬁxed effects estimator (Stock & Watson, 2003). Hausman
tests on all models (insignificant) suggested that random effects regressions were
appropriate for the analyses. Variance inﬂation factors were used to assess the potential
for multicollinearity in the data. Despite signiﬁcant correlations between several
individual measures (e. g. 0.85 between two of the experience indicators and 0.84 between
breadth of opening distribution release and box ofﬁce revenue), calculated variance
inﬂation factors indicated that the collinearity in the data did not reach problematic levels
(Kutner, N achtsheim, and Neter [2003] suggest average VIFs above 4 can be biasing; no
VIF was higher than 2, except in models where all three types of experience were
considered together), indicating that the presence of some collinearity was not enough to
bias the results of the analyses. Still, given the high individual correlations between
variables in certain models, and particularly given the high correlations between
experience variables, alternative speciﬁcations of models are presented in sequential
order to show the effects of different subsets of variables and to investigate the inﬂuence

of multi-collinearity in the data.

111,

.5. v a ancEwi 2a 5..

45:23.3 .33 332:; 035 3 ms .222. 2: 39:28.5 3? 2:: 2 25 _
5:: 53on 33:22-50 .5. v m Eng-tam? 2a 5. SE. .53on maggots-u

 

. I . .. . . - . . . . . ... . £2,333,595
- 3. no 2 xx 3 5 5 _ _ mm 3 x. on c 5 5 mega Gusto o>_§_o.m-ELE .m.
. i. . . . . . ... I . . .. . 3238—005
- 5 mo 3 :m E em .0 no x. mm mm o m. o 350 xom Essex-FEE .:
-- 5 5m: mm 5. 5. 3. 5. 5.- mo. .md _ .5 $539.85 .325 .c_
- cm cc 5.- _m mm 5.- we. 5. x3. .325 5.38.85 quO gem .o
- cm. 5.- :5. o _. m5. 5.- 5.- or... mm.m mwccam 3.52:0 .x
- 5.- 3w. 5. 5. mo. 5. 5N. med. 33:33— 350 ii <-
: 5.- 5. 5.- 5.- 5.- mm... fun £2.th 55.5355 n .c
. .- .- .- $583. 3
- no 5 5 5 55 :2: 333% $5.30 5 5235 .m-
-- x3. 8. cc. xx: 2 .2 28.25 8:395 a.
. . . . Sac—E STE E cacumoaom 3:80
- mm 5 w. v ca v 33:2 50 E 5985 oucotogxm .m
. .- :- :EE Each .3 2:00 25% E one—2
- mw mm m £4 0 8332 SE 3 Saun— uucotoaxm .m
i... -. 3:80 .95 E one—z
- no a. :- w 3332 SE 5 3:23me 5:30 ._
N— Z 3 a w b e m w m N _ $5 .Em 53: 933...;

 

80:28.50 98 mesmcﬁm ozatomoa I m 2an

112

.32: wES: 252.5 035:3: wEmEE mEE _ N. 9 26 223503: 53$ :2 o. swab: c 2252 -

.055 £5 E BEBE 8: 2: 3:06:38 E05 :3 agate note: ousﬁE 30:22 _
.53. vane; ”5. vain ”mo. VA... ”2. VQ+

 

 

 

mwxv mwwv mwxw max? nwmw 0mm 0mm own 0mm 0mm 2
00. cc. cc. mo. no. :m 0o. Oh. cc. «:0. wk
*fwwhﬁm **¢h.mm .13...ch 3&ch tum-6N 5.300: .13.be .I:..hw: ism-w:— iioh: nNECﬁ
3:. _ - 3V. _ - Nm. _ - ﬁn. _ - 0v. _ - ***m-m-.O_ .I:..hm-d— ***Nm.c_ ***Nm.c_ as; _ rem: EEmCOU
. . . . mEzm m-:E.t.¥>..-w
5 - 5 - No 5 :25 :_ 8.8353. \c 55%:
8:60 2.3:: :o n .
- . . . . FEE mist-VSM
M: NO - *3 *mo 133m mm 0.500 053m K: it; ..
P: 031.2 m0_>O—Z hcmhnm t AN
8:00 as: E EEK «NEE-SSH \a
5.- 5.- S.- S. E 833: . .
m0_>cz hora # Tyrj \N3KVLQ
mar NC.- NOE Nor NON _o. _C.- _c.- _O.- _O.- ﬁver—an— EEUEXEQ t
. . . .- .- . . . . . . ammo—ow—
***_C- “In—O- ***_O- xxx—O .I:..—C fix—O tin—O xii—C .I:..—O .I:..—C wﬁcoQO-E 5050.5
tum-4“. .I:.Mv. 3.3;... 3.2.9.. ***Nﬁ. tin—N. ***_N. visa—N. fin—N. flu—N. Eat—gm— cozy-Exzm
** I .3." .I.. «.2. 3.. .3." ** .I.. .I.. _m_o.::oU totem QNQEEQU
A: 352 a Etc—2 w 3662 h 352 e Etc—2 m 38: Q 382 m 322 N 382 — fete—Z 03353» ﬁns-JV

 

 

mass: .855

 

2533— SEC “Sm

 

 

8:§E£Bm 5296 8 8:25me no 5585 5:: £30 :8 m:o_mwo:wom ESE I o 03::-

113

 

6:6 was... :53... 032...: wEmmE. 2...... I. o. 2.6 22.5203». .038 can 2 ﬁne... 0 £2.32 N.
.038 E... E .5..qu .3: 2.. 856508 .3... =5 ﬂout“. cocoa 2.20:. 2252 _

 

 

 

 

._OO.VQ*** u_O.VQ..c.n "no. van." ”2. VQ ...
0mm 0mm 9mm 0mm. 3mm 0mm 9mm 0mm 0mm 0mm 2
«3.3.50 3.3..wa _c Styx??? ***\.c.3m ***3.0M 3.3.59.0 ***NN.Om .139...“ *usiloém .I:..x-ém KENS:
**3.m_- :Nm.m_ - «*NCM..- :5...— _- 515.2- +3.41 .57? +3.47 33.4.- $6.4..- =5...”ch
. . . . £5.”— Etn. E 3.58.5“ m...:.....i...m
11% ***mm c: .c 8.50 2.3:: .o u .6 .352m
- .- . . . EE Boon— mm 250 25% m,..........i,..m~
*5 2m. .22 *c: E 2.32 8.2.2 .25 t x.» $35
.- . .. .- . 8:00 i... E FEE 32......SM?
2. too a. z. 3 33.5.3. 3302 St; a .23 .332.
.8. m... z..- c... 8.- z..- x..- z..- 3.- 8.- 35:... 8.625... a
”IE—C. ***_C. flu—O. "ti—3. .3310. 3.13. “I..—C. “tin—C. ***_C. “tin—O. 9.2.201 %~__:un_o._c zﬁngm
3. R. cm. 3»: c... E. *m. _. m _. m _. 2. .owcsm 5.5.62“.
5.. “I. .2" .3." 3.. “I, .3." .3." “.3 .I.. .m_c.::cU GOT—om £92.59
2 .252 a .232 a .252 h .982 e .252 m 3.52 v .252 n .232 N .252 . .922 2.3.5» 955
maﬁa—5:85 35¢ .25 3353—.553 €333.39:— 856 .8:

 

 

 

85.8.8.9. 850 xom 553.605 9 coautomxm— no 5935 can 530 5. 205$..me 32mm ozﬂwoq I o. 03an

114

 

6:6 mass. .328 032:3“ wEEE mEE F S 26 8235030 838 on: £252 m
.033 £5 E 39:05 8: 2m 353508 :2: :5 58:0 @2qu 25.2: 2252 _
.25. vain.“ Jo. vase.“ HQVQ“: 5.. VQ +

 

 

 

“xx.“ “xxv nxwv nxwv “xxv ”wwv “wwv “wwv “wwv ”mm? 2
***m~.vv .21 _ Wm? trend...” is, 36m *mm.v_ +8.2 *wm.o_ ‘592 10.3 *0w.o_ MN 333
**8.¢_- “312.2- *mo.m_- “1.0.2- *ucﬁdT **wm.m_- #19:- **wm.m._- «10.2- “156.2- ngcoU

. . . . £55 SE E 125353. E§.z.x>.m
.I:."bv ***#N _O _O why—:00 033:: .3 t \c Shaka
.- . .- .- 5:”. .33; 3 2:00 oEam mgzmtmaﬂ

.38 .103 No me E 2.32 8:52 SE t \c SR5

- . . . . 2:00 he... E FEE mgzﬁtﬁm
mc - 33% co co 3 3.323. 8392 SE n \e 3:5 :256
m _. E. as. 2. as. No. me. No. No. 8. E055; 22.2605 .4
CC. 33. CC. C3. 5.. 5.- _o.- 5.- 5.- 5.- ammo—um choao ho Suaohm
F: ww. mm. cm. .I:..“. *5. £3. i3. i3. *5. awasm 5:83—55
“.3." .I.. .3.“ “In .1" .In .I.. .In ** in _m_c._=_oU Coton— 9.3.2239
c— .252 a .282 w .232 b .252 e .252 m .253 v .232 n .252 N .252 _ E52 033...; 9.9.0

 

£238.85 95:. .825 233-5;

 

ESSESE .8:th

 

 

85.50%me wwccmm Bantu Basﬁooa 9 oocotomxm mo .23on can 595 H8 3232QO ESE ocﬁwoq I Z 035%

115

Results

Table 8 provides correlations and descriptive statistics for the variables in the
analyses. Tables 9, 10, and 11 provide the results for hypothesis testing. The analyses
provide mixed support for Hypothesis 1. In fact, the analyses provided no support for the
effects of overall experience on overall performance, in terms of box ofﬁce or critical
ratings. However, there is some evidence that when considered alone, overall experience
does help firms to create ﬁrm-relative blockbusters (see Model 7 on both Tables 10 and
11) both in terms of critical and box ofﬁce success. This effect is overwhelmed by the
effects of experience breadth when considered together. In total, the results provide, at
best, some mixed support for the universalistic perspective. The results also provide
support for Hypothesis 2. Model 8 on Tables 10 and 11 provide support for Hypothesis 2,
suggesting that increasing levels of experience breadth do promote blockbuster
performance relative to a firm’s previous success. Further, in models that consider all
three types of experience, the effect of experience breadth overwhelms the inﬂuences of
any other type of experience, suggesting that breadth really drives the creation of ﬁrm-
relative blockbusters over time (see Model 10 on both Tables 10 and 11). However,
these effects were not found with broad market-based indicators of blockbuster
performance. The results also provide a mix of support for Hypothesis 3. When
considered in isolation, depth was found to enhance the likelihood of blockbuster
performance, in terms of market-based and firm-relative box ofﬁce blockbuster
performance, and in terms of firm-relative critical blockbuster performance. This finding
was opposite to the premise of Hypothesis 3. However, support for Hypothesis 3 can be

found in Model 10 on both Tables 10 and 11, where the effects of experience depth

116

significantly negatively inﬂuence the creation of firm-relative blockbusters when the
effect of experience breadth is controlled in a fully saturated model. This flipped sign
issue points to an important finding that deserves a little extra attention. When considered
in isolation, the independent effects of different types of experience can be significant
and positive (e.g. Model 8 in Tables 10 and 11) but then flip signs and be a signiﬁcant
and negative (e. g. Model 10 in Tables 10 and 11) when considered jointly with other
experience constructs. For example, in Model 8 from Table 10 the effect of experience
depth is positive and significant in predicting box ofﬁce blockbusters, but reverses sign
when its shared variance (its correlation with experience breadth is .75) is partialed out in
a more saturated model and it is left to only explain its unique variance. Thus, when
considered alone, experience depth helps to promote firm-relative blockbusters, but this
effect is largely driven by its proxying effect for experience breadth. When the two are
allowed to account for only their unique variance, the effect of depth actually hurts a
firm’s chances for achieving a blockbuster (as hypothesized in Hypothesis 3). This
situation provides a major caution for the evaluation of the effects of experience in the
broader experiential learning literature, and provides strong evidence for the need to
examine multiple types of experience in research.
DISCUSSION AND CONCLUSION

These results provide some support for both the universal and contingency
perspectives, but overall fairly limited support for the universalistic perspective. It is
simply not clear that any experience is valuable experience. In fact, in many models
experience just did not matter (non-signiﬁcant ﬁndings). Further, in other cases overall

experience did not matter, but more specific types of experience were very signiﬁcant

117

(depth in some cases, breadth in other cases). Thus the conventional ‘common sense’ is
not quite as obvious as it may seem. The implications of this research can be broken
down into three categories.

First, this research has implications for experiential learning theory and research.
Experiential learning research has made signiﬁcant progress in recent years, and the
current ﬁndings help to advance this research in several ways by both providing new
answers and raising new questions. Although the primary contingencies hypothesized
focused only on the type of experience (depth and breadth) and the type of performance
(blockbuster) as contingencies, these results provide strong evidence for the importance
of several contingencies for the relation between organizational experience and
performance. Speciﬁcally, contingencies found to have an inﬂuence on the results
included the type of experience (depth, breadth, and overall), the type of performance
assessment (overall performance vs. blockbuster success), the audience for these
performance evaluations (critical ratings vs. commercial success), and the locus of
performance evaluation (internal reference points or market-based comparisons). Each of
these different contingencies played a role in the results presented here, highlighting the
complex nature of experiential learning in organizations. In fact, these results may point
to a big-picture contingency perspective, where in some cases the universal perspective
may have some merit. However, in most cases, various contingencies inﬂuence the effect
of experience on performance.

Another potential contribution of these findings for experiential learning research
is that these findings suggest that previous research on organizational learning from

experience may be misleading. Most notably, these results suggest that we may want to

118

re-examine prior findings on the relation between organizational experience and
performance. In these data, although an overall measure of experience was signiﬁcant in
some models when considered in isolation, more precise measures of experience
overwhelmed the effects of this generalized measure. Further, in cases where the overall
measure of experience was not signiﬁcant, other more specific measures of experience
were significant. Due to the nature of publishing standards in the organizational sciences,
non-signiﬁcant results often do not get published (Rosenthal, 1979; Sigelman, 1999).
This suggests that re-examining previous research for non-significant ﬁndings that simply
employed generic ‘overall’ experience measures might be fruitful grounds for examining
the effects of more specific measures of organizational experience or different types of
performance evaluations. As ‘overall’ experience measures are common, particularly in
organizational theory and strategy research for which experiential learning is not the
primary focus, this potentially represents a fair number of studies to re-examine.

Second, this research has implications for organizational performance research.
Most notably, these ﬁndings provide evidence that research on organizations should
continue to branch out to consider alternative approaches to performance. As different
approaches to performance have both theoretical (the subjective evaluation of creativity
vs. financial performance) and practical (the need to achieve a blockbuster hit to establish
a reputation or even survive in an industry) roots and implications, there is a need for
more diverse methods to evaluate organizational performance. As was demonstrated in
this research, each of the three different approaches to performance considered here was
found to be uniquely inﬂuenced by experiential learning. These included the type of

performance assessment (overall performance vs. blockbuster success), the audience for

119

these performance evaluations (critical ratings vs. commercial success), and the locus of
performance evaluation (internal reference points or market-based comparisons). Further,
this highlights the need for more research into the nature of creative success, and how
such perceptions are formed. Interestingly, the formation of creativity perceptions has
received little attention in the organizational literature (Amabile, et. al, 1996; Flowers &
Garbin, 1989). Given the importance of critical ratings in many entertainment contexts
(art, music, theatre, etc.) and the increased prominence of design and aesthetics in
product development (Eisenman, 2007; Nussbaum, 2004), this represents an important
line of inquiry for future research.

Third, this research has implications for the industry and for practicing managers.
For the movie industry, these results confirm the folk wisdom that very little helps to
predict what positively inﬂuences critical ratings (see Models 6-10 in Table 9 and Models
1-5 in Table 11). This builds on the earlier point that we simply do not have much
research on this important performance outcome, despite its importance to industry.
However, these results do provide a clearer path for production companies that hope to
improve their own results over time through experience, and provide some insight on
how to create box office blockbusters (via depth and breadth of experience). For the
broader practitioner audience, this research also has several implications. First, it points
to the differing effects of organizational experience depth and organizational experience
breadth. Such differences could have important policy implications for organizations
attempting to become ‘learning organizations’ (Senge, 1990). Further, this research in
some ways refutes the generalized idea that “any experience is good experience”

common sense widespread among executives, as many models had non-significant

results, or even negative relations as the effects of organizational experience. Thus,
corporate executives should be more discriminating in considering organizational
experience, to carefully consider the type of experience that would best suit the
organization and its performance goals.

In conclusion, the goal of this research was to shed new light on the nature of
experiential organizational learning by probing differences between different types of
organizational experience and performance. In doing so, several avenues for future
research have been uncovered. Hopefully this will encourage more research into some of

these complex but interesting questions.

CONCLUSION

In reﬂecting back on the core themes of this dissertation, these three chapters help
to overcome several significant limitations in the traditional organizational experiential
learning literature. As has been a consistent theme throughout this dissertation,
historically research has made generalizations about organizational experience that have
limited the scope of this work. Further, and perhaps more importantly, almost no research
has considered different types of experience simultaneously to disentangle the unique and
joint inﬂuences of different types of experience. This dissertation has helped to address
these shortcomings in several ways.

Chapter 1 developed a new theoretical model of organizational experience to help
disentangle these different types of experience. The key to this effort was bringing
together micro-level research (e. g. Tesluk & Jacobs, 1998) and integrating it with more
recent advances in organizational-level research (e.g. Nerkar & Roberts, 2004), with
Table 1 as the result. This integration was critical to the development of the rest of the
dissertation, where these different types of experience were used to motivate new
directions in several firm processes and outcomes. Continuing on with Chapter 1, this
framework provided a new theoretical approach to the antecedents of search in firms, a
new direction for the line of work launched by Cyert and March (1963). In Chapter 2, this
framework helped to motivate new theoretical directions in the origins of knowledge
creation in firms, responding to calls for such work from Nonaka and colleagues
(Nonaka, von Krogh, & Voelpel, 2006) where different types of experience were
hypothesized (and found) to promote different knowledge creation outcomes. Finally, in

Chapter 3, this new framework provided the base for hypothesizing differential

theoretical links between different types of organizational experience and different firm
outcomes (universal and contingency perspectives), answering the call for such work
from Argote (1999).

In addition to these new theoretical directions, in Chapters 2 and 3 the problems
with prior research approaches to organizational experience were demonstrated
empirically. Without these empirical tests, the premise of this dissertation that different
types of organizational experience have different inﬂuences on ﬁrm processes and
outcomes would simply be a theoretical exercise. With these empirical chapters, these
criticisms are borne out with empirical evidence. In fact, in two very different industrial
settings, different types of organizational experience were found to have very different
relations with knowledge creation and performance. Further, and a more important
critique of the previous literature, is my consistent finding that even when some measures
of experience are not significant, specific types of experience can be signiﬁcant, and have
different signs. This suggests that prior research that has only considered a single type of
experience in a study may not be providing a full picture of the nature of experiential
learning. As was demonstrated in Tables 10 and 11, in many cases when individual
metrics were considered in isolation their effects were dramatically different (in sign and
significance) than when considered jointly with other types of experience together. This
suggests that prior ﬁndings presented in the traditional organizational learning literature
might be worth revisiting to assess whether the found relations were the result of true
findings or proxy effects for an unmeasured dimension of experience. In these cases it is

at least possible that the unique inﬂuences of found relations might be subject to change.

123

In addition to these broad implications for the organizational learning literature,
together the three chapters in this dissertation also provide a wide set of speciﬁc new
findings regarding the nature of organizational experience and learning. Starting in
Chapter 2, there are several ﬁndings regarding the nature of organizational experience
and knowledge creation of note. First is the split finding regarding the inﬂuence of
experience depth on the creation of high impact knowledge. As expected, time-based
approaches were negatively related to impact (core-rigidities). However, quantity-based
approaches were positively related to impact. This raises interesting questions about prior
research, which has alternated between these two measurement approaches (time and
quantity) without much theoretical basis, suggesting notable differences that need to be
explored. In addition, the results regarding experiential breadth are also intriguing, as
they also split a construct that was hypothesized to be uniform. In fact, organizational
breadth (working for a wide variety of companies) and technological breadth (working in
a wide variety of different technical areas) were found to have distinctly different
inﬂuences on knowledge creation in the pharmaceutical industry. Finally, in this chapter
experiential diversity was introduced to consider how mixed teams of scientists with
varying levels of experience were able to work together to create new knowledge.
Although some types of diversity had the expected positive inﬂuence on knowledge
creation with diminishing returns, other types of diversity had negative inﬂuences on the
impact of knowledge creation efforts. Probing further into these differences will provide
new directions for future research.

Chapter 3 also provides evidence for several expected relations together with

several surprising findings. In this case, the relations between different types of

124

organizational experience and performance were found to vary across several different
contingencies, including the type of experience (depth, breadth, and overall), the type of
performance assessment (overall performance vs. blockbuster success), the audience for
these performance evaluations (critical ratings vs. commercial success), and the locus of
performance evaluation (internal reference points or market-based comparisons). A
particular strength of chapter 3 was the evaluation of experiential learning inﬂuences on a
wide variety of related but distinct forms of performance. The fact that these different
types of performance, though related, provided a fair degree of divergence in signiﬁcance
and signs in relations provides a strong argument for a broader approach to the study of
organizational learning in future research.

In addition to these contributions to organizational learning research, this research
provides several findings relevant to organizations and their managers. Most notably is
the potentially differential value of different types of experience for organizations. By
raising serious challenges to a universalistic perspective of ‘any experience is good
experience’ and demonstrating a disparity in the value of different types of experience in
different settings, this research cautions managers in their pursuit of creating learning
organizations. In particular, this research suggests that managers should assess their
vision for organizational learning, and then build experience strategically to support such
learning goals.

In addition to the directions described above, this research also provides
additional new directions for future research to uncover the power of experiential
learning in all of its forms. In particular, there is a signiﬁcant opportunity to explore the

dimensions of organizational experience that diverged in these analyses (time vs.

125

quantity, organization vs. technical breadth, etc.) to provide theoretical direction in these
cases. A starting point may be to look back before looking forward, to re-examine
existing research in the learning curve tradition for evidence of time-based vs. quantity-
based splits in prior research that were potentially disparate and have escaped attention.
In addition to exploring these differences, Chapter 1 in particular provides a big picture
vision for the future of organizational learning research to push further into the
dimensionalization of experience. If the empirical results of Chapter 2 and 3 are any
indication, then further research into the timing, density, complexity, and other
dimensions of experience may hold additional promise in understanding the rich
dynamics of experiential learning in and by organizations. Further, although micro-
mediational mechanisms were largely beyond the scope of this research, the divergence
in signiﬁcant relations between different types of experience may suggest that different
mediational processes may lie between these experiences and various outcomes. By
pressing down into the different inﬂuences of experience on behavior, norms, routines,
and culture in organizations we may find notable differences to support these divergent
findings.

Finally, as has been highlighted throughout the conclusions of the chapters, this
research provides signiﬁcant cautions for the ﬁeld against simple approaches to
experiential learning in organizations. With singular approaches to experience or
unsophisticated measures, there is a strong potential for oversimplification of experiential
learning models into heuristics that may be unpredictable in different settings. This leads
to a real possibility for unreliability in research findings, and even a potential for opposite

significant relationships (due to unmeasured proxy effects). Hopefully, this dissertation

provides caution for the field in the future measurement and conceptualization of

organizational learning research.

127

APPENDIX A: Pharmaceutical Sub-Domains

Industry SIC Codes:

283 Drugs
2833 Medicinal Chemicals and Botanical Products
2834 Pharmaceutical Preparations
2835 In-Vitro and In-Vivo Diagnostic Substances
2836 Biological Products, Except Diagnostic Substances

Corresponding NAICS Cocks;

3254 Pharmaceutical and Medicine Manufacturing
325411 Medicinal and Botanical Manufacturing
325412 Pharmaceutical Preparation Manufacturing
325413 In-Vitro Diagnostic Substance Manufacturing
325414 Biological Product (except Diagnostic) Manufacturing

APPENDIX B: Film Genres

GESEZSOmQQthN—

Adventure
Animated
Comedy
Crime
Drama
Fantasy
Historical
Horror
Musical
Mystery
Romance
Sci-Fic
Thriller
War
Western

129

REFERENCES

Adner, R., & Levinthal, D. (2001). Demand heterogeneity and Technology Evolution:
Implications for Product and Process Innovation. Management Science, 47, 611-628.

Agrawal, A., & Henderson, R. (2002). Putting patents in context: Exploring knowledge
transfer from MIT. Management Science, 48, 44-60.

Ahuja, G. (2000). Collaboration networks, structural holes, and innovation: A
longitudinal study. Administrative Science Quarterly, 45, 425-445.

Ahuja, G., & Katila, R. (2001). Technological acquisitions and the innovation
performance of acquiring firms: A longitudinal study. Strategic Management Journal,
22, 197-220.

Ahuja, G., & Lampert, C. M. (2001). Entrepreneurship in the large corporation: A
longitudinal study of how established ﬁrms create breakthrough inventions. Strategic
Management Journal, 22, 521-543.

Alchian, A. (1963). Reliability of progress curves in airframe production. Econometrica,
31, 679-693.

Allen, T. J. (1977). Managing the ﬂow of technology. Cambridge: MIT Press.

Amabile, T. M, Conti, R, Coon, H, Lazenby, J, Herron, M. (1996), Assessing the work
environment for creativity, Academy of Management Joumal, 39, 1154-84.

Amason, A. C. (1996). Distinguishing the effects of functional and dysfunctional conﬂict
on strategic decision making: Resolving a paradox for top management teams.
Academy of Management Journal, 39, 123—148.

Amburgey, T. L., Kelly, D., & Barnett, W. P. (1993). Resetting the clock - The dynamics
of organizational-change and failure. Administrative Science Quarterly, 38, 51-73.

Argote, L. (1993). Group and organizational learning curves: individual, system and
environmental components. British Joumal of Social Psychology. 32, 31-51.

Argote, L. (1999). Organizational Learning: C reating, Retaining and Transferring
Knowledge. Boston, MA: Kluwer Academic Publishers.

Argote, L., & Darr, E. (2000). Repositories of knowledge in franchise organizations:
Individual, structural, and technological. G. Dosi, R. R. Nelson, S. G. Winter, eds.
The Nature and Dynamics of Organizational Capabilities. Oxford University Press,
New York.

130

Argote, L., & Epple, D. (1990). Learning curves in manufacturing. Science, 247, 920-
924.

Argote, L., Beckman, S. L., & Epple, D. (1990). The persistence and transfer of learning
in industrial settings. Management Science, 36, 140-154.

Argyris, C. & Schon, D. A. (1978). Organizational learning: A theory of action
perspective. Reading, MA: Addison-Wesley Publishing Company.

Asher, H. (1961). Cost and quantity relationships in the airframe industry. Report R-291,
RAND Corporation.

Austin, J. R. (2003). Transactive memory in organizational groups: The effects of

content, consensus, specialization, and accuracy on group performance. Journal of
Applied Psychology, 88, 866-78.

Baloff, N. (1966). Learning curve: Some controversial issues. Journal oflndustrial
Economics, 14, 275-282.

Baloff, N. (1971). Extension of learning curve: Some empirical results. Operational
Research Quarterly, 22, 329-340.

Bapuji, H., & Crossan, M. (2004). From questions to answers: Reviewing organizational
learning research. Management Learning, 35, 397-417.

Barkema, H. G., Shenkar, O., Vermeulen, F., & Bell, J. H. (1997). Working abroad,
working with others: How ﬁrms learn to operate joint ventures. Academy of
Management Journal, 40, 426-442.

Barney, J. B. (1991). Firm resources and sustained competitive advantage. Journal of
Management, 1 7, 99- 120.

Barrett, F. (1998). Creativity and improvisation in jazz and organizations: Implications
for organizational learning. Organization Science, 9, 605-622.

Baum, J. A. C., & Ingram, P. (1998). Survival-enhancing learning in the Manhattan hotel
industry, 1898-1980. Management Science, 44, 996-1016.

Baum, J. A. C., Li, S. X., & Usher, J. M. (2000). Making the next move: How
experiential and vicarious learning shape the locations of chains' acquisitions.

Administrative Science Quarterly, 45, 766-801.

Bettenhausen, K., & Mumighan, J. K. (1985). The Emergence of Norms in Competitive
Decision-Making Groups. Administrative Science Quarterly, 30, 350-372.

131

Bierly, P., & Chakrabarti, A. (1996). Generic knowledge strategies in the US.
pharmaceutical industry. Strategic Management Journal, 1 7, 123-135.

Blau, J. R., McKinley, W. (1979). Ideas, complexity, and innovation. Administrative
Science Quarterly, 24, 200-219.

Boland, R .J. J ., Tenaski, R. V., & Te’eni, D. (1994). Designing information technology
to support distributed cognition. Organization Science, 5, 456-47.

Brown, J. S. & Duguid, P. (1991). Organizational learning and communities of practice:
Toward a unified view of working, learning, and innovation. Organization Science, 2,
40-57.

Burkhardt, M. E., & Brass, D. J. (1990). Changing patterns or patterns of change: The
effects of change in technology on social network structure and power. Administrative
Science Quarterly, 35, 104-127.

Camerer, C., & Vepsalainen, A. (1988). The economic-efficiency of corporate culture.
Strategic Management Journal, 9, 115- 126.

Cameron, A. C. & Trivedi, P. K. (1998). Regression Analysis of Count Data. Cambridge:
Cambridge University Press.

Chao, G. T., O’Leary-Kelly, A. M., Wolf, S., Klein, H. J ., & Gardner, P. D. (1994).
Organizational socialization: It’s content and consequences. Journal of Applied
Psychology, 79, 730-743.

Compeau, D. R., Higgins, C. A., & Huff, S. (1999). Social cognitive theory and
individual reactions to computing technology: A longitudinal study. MIS Quarterly,
23, 145-158.

Conway, R., & Shultz, A. (1959). The manufacturing progress function. Journal of
Industrial Engineering, 10, 39-53.

Cook, S. D. N., & Yannow, D. (1993). Culture and organizational. learning. Journal of
Management Inquiry, 2, 373-390.

Crossan, M. M., Lane, H. W., & White, R. E. (1999). An organizational learning

framework: From intuition to institution. Academy of Management Review, 24, 522-
537.

Cyert, R. M., & March, J. G. (1963). A behavioral theory of the ﬁrm. Blackwell:
Cambridge, MA.

Daft, R. L., & Weick, K. E. (1984). Toward a model of organizations as interpretation
systems. Academy of Management Review, 9, 284-295.

132

Dahlin, K., & Behrens, D. (2005). When is an invention really radical? Deﬁning and
measuring technological radicalness. Research Policy, 34, 717-737.

Darr, E. D., Argote, L., & Epple, D. (1995). The acquisition, transfer, and depreciation of
knowledge in service organizations: Productivity in franchises. Management Science,
41, 1750-1762.

Davenport, T. H., & Prusak, L. (1998). Working knowledge: How organizations know
what they know. Cambridge, MA: Harvard Business School Press.

Davis, K. (1950). Human Society. New York: Macmillan.

De Dreu, C. K. W., & Weingart, L. R. (2003). Task versus relationship conﬂict, team

performance, and team member satisfaction: A meta-analysis. Journal of Applied
Psychology, 88, 741-749.

Dess, G., & Robinson, R. (1984). Measuring Organizational Performance in the Absence
of Objective Measures: The Case of the Privately-Held Firm and Conglomerate

Business Unit, Strategic Management Journal, 5, 265—273.

Dick, A. R. (1991). Learning by doing and dumping in the semiconductor industry.
Journal of Law & Economics, 34, 133-159.

Dodgson, M. (1991). The management of technological learning. Berlin: De Gruyter.

Dodgson, M. (1993). Organizational learning: Review of some literature. Organization
Studies, 14, 375-94.

Dougherty, D., & Heller, T. (1994). The illegitimacy of successful product innovation in
established firms. Organization Science, 5, 200—218

Downs, G. W., & Mohr, L. B. (1976). Conceptual issues in the study of innovation.
Administrative Science Quarterly, 21, 700-714.

Dutton, J. M., Thomas, A., & Butler, J. E. (1984). The history of progress functions as a
managerial technology. Business History Review, 58, 204-233.

Easterby-Smith, M. (1997). Disciplines of organizational learning: Contributions and
critiques. Human Relations, 50, 1085-1113.

Eisenhardt, K. M., Kahwajy, J. L., & Bourgeois 111, L. J. (1997). How management teams
can have a good ﬁght. Harvard Business Review, 75, 77-85.

Eisenman, M. (2007). Cultivating Taste: Constructing the Importance of Product

Aesthetics in Technological Industries. In George T. Solomon (Ed.), Proceedings of
the Sixty-Sixth Annual Meeting of the Academy of Management, ISSN 1543-8643.

133

Epple, D., Argote, L., & Murphy, K. (1996). An empirical investigation of the
microstructure of knowledge acquisition and transfer through learning by doing.
Operations Research, 44, 77-86.

Ettlie, J. E., Bridges, W. P., & O’Keefe, R. D. (1984). Organization strategy and
structural differences for radical versus incremental innovation. Management Science,
30, 682-695.

Feldman, D. C. (1981). The multiple socialization of organization members. Academy of
Management Review, 6, 308-318.

Feldman, D. C. (1984). The development and enforcement of group norms. Academy of
Management Review, 9, 47-53.

Feldman, M. S., & Pentland, B. T. (2003). Re-conceptualizing organizational routines as
a source of ﬂexibility and change. Administrative Science Quarterly, 48, 94-118.

Ferriani, S., Cattani, G., Baden-Fuller, C. (2007). Fitness determinants in creative
industries: A longitudinal study on the ﬁlmmaking industry 1992-2003. In F. Malerba
and U. Cantner (eds.), Innovation, Industrial Dynamics and Structural
Transformation: Schumpeterian Legacies, Springer.

Fiol, C. M. (1994). Consensus, diversity, and organizational learning. Organization
Science, 5, 403-420.

Fiol, C. M., & Lyles, M. A. (1985). Organizational learning. Academy of Management
Review, 10, 803-813.

Fleming, L. (2001). Recombinant uncertainty in technological search. Management
Science, 47, 117-132.

Fleming, L. (2002). Finding the organizational sources of technological breakthroughs:
the story of Hewlett-Packard's thermal ink-jet. Industrial and Corporate C hange, I I,
1059- 1084.

Fleming, L., & Sorenson, O. (2004). Science as a map in technological search. Strategic
Management Journal, 25, 909-928.

Flowers, J. H., & Garbin, C. P. (1989). Creativity and perception. In Handbook of

creativity, ed. J. A. Glover, R. R. Ronning, and C. R. Reynolds, 147-62. New York:
Plenum.

Ford, J. K., Quinones, M. A., Sego, D. J ., & Sorra, J. S. (1992). Factors affecting the
opportunity to perform trained tasks on the job. Personnel Psychology, 45 , 511-527.

134

Francis, J. N. P. (1991). When in Rome? The effects of cultural adaptation on

intercultural business negotiations. Journal of International Business Studies, 22,
403-428.

Friedrich, R. J. (1982). In defense of multiplicative terms in multiple regression
equations. American Journal of Political Science, 26, 797-833.

Gans, J. S., & Stern, S. (2003). The product market and the market for “ideas”:

commercialization strategies for technology entrepreneurs. Research Policy, 32, 333-
350.

Garcia, R., & Calantone, R. (2002). A critical look at technological innovation typology
and innovativeness terminology: a literature review. Journal of Product Innovation
Management, 19, 110-132.

Garg, A., & Milliman, P. (1961). The aircraft progress curve: Modiﬁed for design
changes. Joumal of Industrial Engineering, 12, 23-27.

Gavetti, G., & Levinthal, D. (2000). Looking forward and looking backward: Cognitive
and experiential search. Administrative Science Quarterly, 45, 113-137.

Gordon, R. A. ( 1968). Issues in Multiple Regression. American Journal of Sociology, 73,
592-616.

Grant, R. (1996). Toward a knowledge-based theory of the firm. Strategic Management
Joumal, 17, 109-122.

Greene, W. H. (1994). Accounting for excess zeros and sample selection in poisson and
negative binomial regression models. Working paper, Stern School of Business, NYU
EC-94-10.

Greene, W. H. (2003). Econometric Analysis, 5th Edition, Prentice Hall.

Gupta, A. K., Smith, K. G., & Shalley, C. E. (2006). The interplay between exploration
and exploitation. Academy of Management Journal, 49, 693-706.

Haleblian, J ., & Finkelstein, S. (1999). The inﬂuence of organizational acquisition
experience on acquisition performance: A behavioral learning perspective.
Administrative Science Quarterly, 44, 29-56.

Haleblian, J ., Kim, J ., & Rajagopalan, N. (2006). Routine-based learning, performance-

based learning, and their interaction: Evidence from the US. commercial banking
industry. Academy of Management Journal, 49, 357-370.

135

Hall, B. H., J affe, A., & Tratjenberg, M. (2001). The NBER patent citation data file:
Lessons, insights and methodological tools. NBER Working Paper #8498.
Cambridge, MA.

Hambrick, R. C., Geletkanycz, M. A., & Fredrickson, J. W. (1993). Top executive
commitment to the status quo: Some tests of its determinants. Strategic Management
Journal, 14, 401-419.

Hannan, M. T., & Freeman, J. (1984). Structural inertia and organizational change.
American Sociological Review, 49, 149-164.

Hargadom, A., & Sutton, RI. (1997). Brainstorming groups in context: Effectiveness in a
product design ﬁrm, Administrative Science Quarterly, 41 , 685-718.

Harrison, D. A., & Klein, K. J. (2007). What's the difference? Diversity constructs as
separation, variety, or disparity in organizations. Academy of Management Review,
32, 1199-1228.

Harrison, J. R., & Carroll, G. R. (2006). Culture and demography in organizations.
Princeton University Press, Princeton.

Hatch, N. W., & Mowery, D. C. (1998). Process innovation and learning by doing in
semiconductor manufacturing. Management Science, 44, 1461-1477.

Haunschild, P. R., & Sullivan, B.N. (2002), Learning from complexity: effects of prior
accidents and incidents on airlines' learning, Administrative Science Quarterly, 47,
609-43.

Hausman, J. A. ( 1978). Specification tests in econometrics. Econometrica, 46, 1251-
1271.

Hausman, J. A., Hall, B. H., & Griliches, Z. (1984). Econometric models for count data
with an application to the patents-R&D relationship. Econometrica, 52, 908-938.

Hedberg, B. (1981). How organizations learn and unleam. In P. C. N ystrom, W. H.
Starbuck, (eds.) Handbook of Organizational Design. Oxford University Press: New
York, 3—27.

Henderson, R. M., & Clark, K. B. (1990). Architectural innovation: The reconﬁguration
of existing systems and the failure of established firms. Administrative Science

Quarterly, 35, 9-30.

Henderson, R., & Cockbum, I. (1994). Measuring competence: Exploring firm effects in
pharmaceutical research. Strategic Management Journal, 15, 63-84.

136

Hirsch, W. ( 1952). Manufacturing progress functions. Review of Economics and
Statistics, 34, 143-55.

Hoang, H., & Rothaermel, F. T. (2005). The effect of general and partner-speciﬁc
alliance experience on joint R&D project performance. Academy of Management
Journal, 48, 332-345.

Hollander, S. (1965). The Sources of Increased Efficiency: A study of the DuPont Rayon
Plants. MIT Press: Cambridge, MA.

Holmqvist, M. (2004). Experiential learning processes of exploitation and exploration
within and between organizations: An empirical study of product development.
Organization Science, 15, 70-81.

Huber, G. P. (1991). Organizational learning: The contributing processes and literatures.
Organization Science, 2, 88-115.

Ichijo, K., & Nonaka, I. (2006). Knowledge Creation and Management: New Challenges
for Managers. Oxford: Oxford University Press.

J aeger, A. M. (1983). The transfer of organizational culture overseas: An approach to
control in the multinational corporation. Joumal of International Business Studies,
14, 91-114.

Jana, R. (2007). Service innovation: The next big thing. BusinessWeek. March 29‘“,
2007.

Jansen, E., & Vonglinow, M. A. (1985). Ethical ambivalence and organizational reward
systems. Academy of Management Review, 10, 814-822.

J ehn, K. A. (1995). A multi-method examination of the beneﬁts and detriments of intra-
group conﬂict. Administrative Science Quarterly. 40, 256—282.

Jehn, K. A. (1997). A qualitative analysis of conﬂict types and dimensions in
organizational groups. Administrative Science Quarterly, 42, 530—557.

J ehn, K. A., Northcraft, G. B., & Neale, M. A. (1999). Why differences make a
difference: A ﬁeld study of diversity, conﬂict, and performance in workgroups.
Administrative Science Quarterly, 44, 741—763.

Katila, R. (2002). New product search overtime: Past ideas in their prime? Academy of
Management Journal, 45, 995-1010.

Katila, R., & Ahuja, G. (2002). Something old, something new: A longitudinal study of
search behavior and new product introduction. Academy of Management Journal, 45 ,
1 183-1194.

137

Katz, D., & Kahn, R. L. (1978). The Social Psychology of Organizations. New York:
Wiley.

Katz, R., & Allen, A. (1982). Investigating the Not Invented Here (NIH) Syndrome: a
look at the performance, tenure and communication patterns of 50 R&D project
groups. R&D Management, 12, 7-19.

Katz, R., & Tushman, M. (1979). Communication Patterns, Project Performance, and
Task Characteristics: Empirical-Evaluation and Integration in an R&D Setting.
Organizational Behavior and Human Performance, 23, 139-162.

Keller, K. L. & Staelin, R. (1987). Effects of quality and quantity of information on
decision effectiveness. Journal of Consumer Research, 14, 200-214.

Kelly, R., & Caplan, J. (1993). How Bell Labs creates star performers. Harvard Business
Review, 71, 128-139.

Kelsey, S. F., Mullin, S. M., Detre, K. M., Mitchell, H., Cowley, M. J ., Gruentzig, A. R.,
Kent, K. M. (1984). Effect of investigator experience on percutaneous transluminal
coronary angioplasty. American Joumal of Cardiology, 53, 56C—64C.

King, A. A., & Tucci, C. L. (2002). Incumbent entry into new market niches: The role of
experience and managerial choice in the creation of dynamic capabilities.

Management Science, 48, 171-186.

Knorr-Centina, K. D. (1981). The manufacture of knowledge: An essay on the
constructivist and contextual nature of science. New York, NY: Pergamon Press.

Kolb, D. (1984). Experiential learning: Experience as the source of learning and
development. Englewood Cliffs, NJ: Prentice-Hall Inc

Kutner M. H., Nachtsheim C. J ., & Neter J. (2003). Applied Linear Regression Models.
4‘h edition, McGraw-Hill Irwing, Boston.

Lawrence, P. R., & Dyer, D. (1983). Renewing American Industry: Organizingfor
Eﬂiciency and Innovation. New York: The Free Press.

Leonard-Barton, D. (1992). Core Capabilities and Core Rigidities - a Paradox in
Managing New Product Development. Strategic Management Journal, I 3, 111-125.

Leonard-Barton, D. (1995). Wellsprings of knowledge: building and sustaining the
sources of innovation. Harvard Business School Press: Boston.

Levinthal, D. (1997). Adaptation on rugged landscapes. Management Science, 43, 934-
950.

138

Levinthal, D., & March, J. (1993). The Myopia of Learning. Strategic Management
Joumal, 14, 95-112.

Levitt, B., & March, J. G. (1988). Organizational learning. Annual Review of Sociology,
14, 319-340.

Lieberman, M. B., & Montgomery, D. (1988). First mover advantages. Strategic
Management Journal, 9, 41-58.

Long, J. S. (1997). Regression Models for Categorical and Limited Dependent Variables.
Thousand Oaks, CA: Sage.

March, J. G. (1991). Exploration and exploitation in organizational learning.
Organization Science, 2, 71-87.

March, J. G., & Simon, H. A. (1958). Organizations. New York, NY: Wiley.

March, J. G., & Sutton, R. I. (1997). Organizational performance as a dependent variable.
Organization Science, 6, 698-706.

Martin, J. (2002). Organizational Culture: Mapping the Terrain. Newbury Park, CA:
Sage.

Martin, J ., & Siehl, C. (1983). Organizational culture and counter-culture: An uneasy
symbiosis. Organizational Dynamics, 12, 52-64.

Martin, X., & Mitchell, W. (1998). The inﬂuence of local search and performance

heuristics on new design introduction in a new product market. Research Policy, 26,
753-77 1.

Martin, X., Swaminathan, A., & Mitchell, W. (1998). Organizational evolution in the
inter-organizational environment: Incentives and constraints on international
expansion strategy. Administrative Science Quarterly, 43, 566-601.

McDaniel, M. A., Schmidt, F. L., & Hunter, J. E. (1988). Job experience correlates of
job-performance. Journal of Applied Psychology, 73, 327-330.

McGregor, J. (2008). The World's Most Innovative Companies. BusinessWeek, April 17,
2008.

McNamara, G., & Bromiley, P. (1997). Decision making in an organizational setting:
Cognitive and organizational inﬂuences on risk assessment in commercial lending.
Academy of Management Journal, 40, 1063-1088.

139

Meyer, J. W., Scott, W. R., & Deal, T. E. (1981). Institutional and technical sources of
organizational structure. In Herman Stein (ed.) Organization and the Human
Services: C ross-Disciplinary Reﬂections: 151-179, Philadelphia: Temple University
Press.

Miller, D. (1996). A preliminary typology of organizational learning: Synthesizing the
literature. Journal of Management, 22, 485-505.

Miller, D., & Shamsie, J. (1996). The resource-based view of the ﬁrm in two
environments: The Hollywood ﬁlm studios from 1936—1965. Academy of
Management Journal. 39, 519-543.

Mowery, D. C., Oxley, J. E., & Silverman, B. S. (1996). Strategic alliances and inter-firm
knowledge transfer. Strategic Management Journal, 17, 77-91.

Muth, J. F. (1986). Search Theory and the Manufacturing Progress Function.
Management Science, 32, 948-962.

Nadler, G., & Smith, W. D. (1963). Manufacturing progress functions for types of
processes. International Journal of Production Research, 12, 115-135.

Nelson, R. R., & Winter, S. G. (1982). An Evolutionary Theory of Economic Change.
Cambridge, Mass: Harvard University Press.

Nemeth, C. J. (1986). Differential contributions of majority and minority inﬂuence.
Psychological Reports, 93, 23-32.

Nerkar, A. (2003). Old is gold? The value of temporal exploration in the creation of new
knowledge. Management Science, 49, 211-229.

Nerkar, A., & Roberts, P. W. (2004). Technological and product-market experience and
the success of new product introductions in the pharmaceutical industry. Strategic

Management Journal, 25, 779-799.

Nonaka, I. (1994). A dynamic theory of organizational knowledge creation. Organization
Science, 5, 14-37.

Nonaka, I., Takeuchi, H. (1995). The Knowledge Creating Company: How Japanese
Companies Create the Dynamics of Innovation. Oxford University Press: Oxford.

Nonaka, 1., von Krogh, G., & Voelpel, S. (2006). Organizational knowledge creation
theory: Evolutionary paths and future advances. Organization Studies, 27, 1179—
1208.

Nussbaum, B. (2004). The power of design. BusinessWeek, May 17, 2004 issue.

140

O'Reilly III, C. A., Chatman, J ., Caldwell, D. F. (1991). People and organizational

culture: Assessing person-organizational fit. Academy of Management Journal, 34,
487-517.

Orlikowski, W. J. (2002). Knowing in practice: Enacting a collective capability in
distributed organizing. Organization Science, 13, 249-273.

Patchen, M. (1962). Supervisory methods and group-performance norms. Administrative
Science Quarterly, 7, 275-294.

Pelled, L. H., Eisenhardt, K., & Xin, K. R. (1999). Exploring the black box: an analysis

of work group diversity, conﬂict, and performance. Administrative Science Quarterly,
44, 1-28.

Pentland, B. T., & Rueter, H. H. (1994). Organizational routines as grammars of action.
Administrative Science Quarterly, 39, 484-510.

Peteraf, M. (1993). The cornerstones of competitive advantage: a resource-based view.
Strategic Management Journal, 14, 179-191.

Pitcher, P., & Smith, A. D. (2001). Top management team heterogeneity: Personality,
power, and proxies. Organization Science, 12, 1-18.

Polanyi, M. (1967). The Tacit Dimension. Garden City, NY: Anchor Books.

Policastro, E. (1995). Creative intuition: an integrative review. Creative Research
Journal, 8, 99-113.

Porter, M. E. (1980). Competitive Strategy. The Free Press: New York.
Popper, M., & Lipshitz, R. (1998). Organizational learning mechanisms: A structural and
cultural approach to organizational learning. The Journal of Applied Behavioral

Science, 34, 161-79.

Popper, M., & Lipshitz, R. (2000). Organizational learning: Mechanisms, culture, and
feasibility. Management Learning, 31, 181-196.

Prahalad, C. K., & Bettis, R. A. (1986). The dominant logic: A new linkage between
diversity and performance. Strategic Management Journal, 7, 485-501.

Preston, L. E., & Keachie, E. C. (1964). Cost-functions and progress functions: An
integration. American Economic Review, 54, 100- 107.

Pugh, D., Hickson, D., Hinings, R., & Turner, C. (1968). Dimensions of organization
structure. Administrative Science Quarterly, 13, 65—91.

141

Quinones, M. A., Ford, J. K., & Teachout, M. S. (1995). The relationship between work
experience and job performance: A conceptual and meta-analytic review. Personnel
Psychology, 48, 887-910.

Raelin, J. A. (1997). Action learning and action science: Are they different?
Organizational Dynamics, 26, 21-34.

Rivkin, J. W. (2000). Imitation of complex strategies. Management Science, 46, 824-844.

Rohrbach, A. (1927). Economical production of all-metal airplanes and sea planes.
Journal of the Society of Automotive Engineers, 20, 57-66.

Rosenkopf, L., & Almeida, P. (2003). Overcoming local search through alliances and
mobility. Management Science, 49, 751-766.

Rosenkopf, L., & Nerkar, A. (2001). Beyond local search: Boundary-spanning,
exploration, and impact in the optical disk industry. Strategic Management Journal,
22, 287-306.

Rosenthal, R. (1979). The “ﬁle drawer problem” and tolerance for null results.
Psychological Bulletin, 85, 63 8-641.

Rothaermel, F. T., & Deeds, D. L. (2004). Exploration and exploitation alliances in

biotechnology: A system of new product development. Strategic Management
Journal, 25, 201-221.

Sampson, R. C. (2005). Experience effects and collaborative returns in R&D alliances.
Strategic Management Journal, 26, 1009-1031.

Scherer, F. M., & Harhoff, D. (2000). Technology policy for a world of skew-distributed
outcomes. Research Policy, 29, 559-566.

Schilling, M. A., Vidal, P., Ployhart, R. E., & Marangoni, A. (2003). Learning by doing
something else: Variation, relatedness, and the learning curve. Management Science,
49, 39-56.

Schmidt, F. L., Hunter, J. E., & Outerbridge, A. N. (1986). Impact of job experience and
ability on job knowledge, work sample performance, and supervisory ratings of job-

perforrnance. Joumal of Applied Psychology, 71, 432-439.

Schwandt, D. R., & Marquardt, M. J. (2000). Organizational Learning: From World-
C lass Theories to Global Best Practices. CRC Press: Boca Raton, FL.

142

Schweiger, D., Sandberg, W., & Regan, J. (1986). Group approaches for improving
strategic decision making: a comparative analysis of dialectical inquiry, devil‘s
advocacy, and consensus approaches to strategic decision making. Academy of
Management Journal, 29, 51-71.

Schwenk, C. (1990). Conﬂict in organizational decision making: an exploratory study of
its effects in for-proﬁt and non-for-profit organizations. Management Science, 36,
436—448.

Scott, W. R. (1995). Institutions and Organizations. Thousand Oaks, CA, Sage.

Selznick, P. (1957). Leadership in Administration. New York: Harper and Row.

Senge, P. M. (1990). The Fifth Discipline: The Art and Practice of the Learning
Organization. Doubleday Currency: New York, NY.

Shane, S. (2001). Technological opportunities and new ﬁrm creation. Management
Science, 47, 205-220.

Sherif, M. (1936). The Psychology ofSocial Norms. New York: Harper & Row.
Shields, C. (2005). Personal communication with Hallmark creative artist.
Sigelman, L. (1999). Publication Bias Reconsidered. Political Analysis, 8, 201-210.

Simonin, B. L. (1997). The importance of collaborative know-how: An empirical test of
the learning organization. Academy of Management Joumal, 40, 1 150-1174.

Simons, T., Pelled, L. H., & Smith, K. A. (1999). Making use of difference: Diversity,
debate, and decision comprehensiveness in top management teams. Academy of
Management Journal, 42, 662—673.

Simonton, D. K. (2003). Scientiﬁc creativity as constrained stochastic behavior: The
integration of product, person, and process perspectives. Psychological Bulletin. 129,

475-494.

Snow, C. C., & Hrebiniak, L. G. (1980). Strategy, distinctive competence, and
organizational persistence. Administrative Science Quarterly, 25, 317-336.

Sorenson, O., Rivkin, J. W., & Fleming, L. (2006). Complexity, networks and knowledge
ﬂow. Research Policy, 35, 994-1017.

Spender, J. C. (1996). Making knowledge the basis of a dynamic theory of the firm.
Strategic Management Journal, I 7, 45-62.

143

 

Stock , J. H., & Watson, M. W. (2003). Introduction to econometrics. Boston, MA:
Addison Wesley.

Stringer, J. (2003). Movie Blockbusters. Routledge: New York, NY.

Stuart, T. E., & Podolny, J. M. (1996). Local search and the evolution of technological
capabilities. Strategic Management Journal, 1 7, 21-3 8.

Sundgren, M., Selart, M., Ingelgard, A., & Bengtson, C. (2005). Dialogue-based
evaluation as a creative climate indicator: Evidence from the pharmaceutical industry.
Creativity and Innovation Management, 14, 84-98.

Tate, R. L. (1984). Limitations of centering for interactive models. Sociological Methods
and Research, 13, 251-271.

Tesluk, P., & Jacobs, R. (1998). Toward an integrated model of work experience.
Personnel Psychology, 51, 321-355.

Utterback, J. M., & Suarez, F. F. (1993). Innovation, competition, and industry structure.
Research Policy, 22, 1-21.

Van de Vliert, E., & De Dreu, C. K. W. (1994). Optimizing performance by stimulating
conﬂict. International Journal of Conﬂict Management, 5, 211-222.

Vandenberghe, C. (1999). Organization culture, person-organization ﬁt, and turnover: A
replication in the health care industry. Journal of Organizational Behavior, 20, 175-
184.

Vaughan, D. (1999). The dark side of organizations: Mistake, misconduct, and disaster.
Annual Review of Sociology, 25, 271-305.

Vera, D., & Crossan, M. (2004). Strategic leadership and organizational learning.
Academy of Management Review, 29, 222-240.

Verona, G., & Ravasi, D. (2003). Unbundling dynamic capabilities: An exploratory study
of continuous product innovation. Industrial and Corporate Change, 12, 577-606.

von Krogh, G., Ichizo, K., & Nonaka, I. (2000). Enabling Knowledge Creation. Oxford
University Press: Oxford.

von Krogh, G., Nonaka, 1., Aben, M. (2001). Making the most of your company's
knowledge: A strategic framework. Long Range Planning, 34, 421-439.

Vroom, V. H. (1960). Some personality determinants of the effects of participation.
Englewood Cliffs, N.J.: Prentice-Hall.

144

 

 

Vuong, Q. H. (1989). Likelihood ratio tests for model selection and non-nested
hypotheses. Econometrica, 57, 307-333.

Walsh, J. P. (1995). Managerial and organizational cognition: Notes from a trip down
memory lane. Organization Science, 6, 280-321.

Walsh, J. P., & Ungson, G. R. (1991). Organizational memory. Academy ofManagement
Review, 16, 57-91.

Wanous, J. P., & Youtz, M. A. (1986). Solution diversity and the quality of group
decisions. Academy of Management Journal. 29, 149-158.

Wegner D. M. (1986). Transactive memory: A contemporary analysis of the group mind.
In B. Mullen & G. R. Goethals (eds.), Theories of group behavior. New York:
Springer-Verlag.

Westphal, J. D., Gulati, R., & Shortell, S. M. (1997). Customization or conformity? An
institutional and network perspective on the content and consequences of TQM
adoption. Administrative Science Quarterly, 42, 366-394.

Winter, S. G. (1994). Organizing for continuous improvement: Evolutionary theory meets
the quality revolution. In Baum & Singh (eds.) Evolutionary Dynamics of
Organizations, 90-108.

Wright, T. P. (1936). Factors affecting the cost of airplanes. Journal of the Aeronautical
Sciences, 3, 122-128.

Yelle, L. E. (1979). The learning curve: Historical review and comprehensive survey.
Decision Sciences, 10, 302-328.

Yukl, G. (1998). Leadership in organizations. Englewood Cliffs, NJ: Prentice-Hall.

Zimmerman, M. B. (1982). Learning effects and the commercialization of new
technologies. Bell Journal of Economics, 13, 297-310.

Zollo, M., & Singh, H. (2004). Deliberate learning in corporate acquisitions: Post-
acquisition strategies and integration capability in US bank mergers. Strategic
Management Journal, 25, 1233-1256.

145

 

   

Ililtil‘lljijllil’ljjﬂl11114111“I