Champions of Product Innovations - Defining, Developing, and Validating A Measure of Champion Behavior

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 22

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/223761301

Champions of Product Innovations: Defining, Developing, and


Validating a Measure of Champion Behavior

Article  in  Journal of Business Venturing · August 1998


DOI: 10.1016/j.jbusvent.2004.06.001

CITATIONS READS
333 3,455

3 authors, including:

Christine Shea
University of New Hampshire
29 PUBLICATIONS   1,189 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Nanotechnology as GPT View project

The Effect of Leadership on Manufacturing Performance View project

All content following this page was uploaded by Christine Shea on 06 October 2017.

The user has requested enhancement of the downloaded file.


Journal of Business Venturing 20 (2005) 641 – 661

Champions of product innovations: defining, developing,


and validating a measure of champion behaviorB
Jane M. Howella,*, Christine M. Sheab,1, Christopher A. Higginsa,2
a
Richard Ivey School of Business, The University of Western Ontario, London, ON, Canada N6A 3K7
b
The Whittemore School of Business and Economics, University of New Hampshire, Durham,
NH 03824-3593, United States
Received 30 July 2002; received in revised form 30 June 2004; accepted 30 June 2004

Abstract

This research defined, developed, and validated a measure of champion behavior by collecting data
from different samples in multiple stages. Using the act frequency method, a comprehensive set of 102
items reflective of champion behavior was developed based on a survey of middle managers and
executives. Next, a different managerial sample rated the extent to which each of the 102 items was
representative of champion behavior, and 29 were deemed to represent the core of the domain of
championship. Finally, to evaluate the psychometric properties of the champion behavior measure, 47
product innovations in 13 firms were studied. Forty-seven nominated champions completed
personality measures, 47 top division managers rated project performance, and 216 innovation team
members rated champion behavior and leadership style. Results from principal components and
confirmatory factor analyses yielded a 14-item champion behavior measure composed of three factors:
expressing enthusiasm and confidence about the success of the innovation, persisting under adversity,
and getting the right people involved. This measure showed acceptable reliability as well as convergent

B
This research investigation was generously supported by a Social Sciences and Humanities Research Council
of Canada Grant (# 410-98-0373) awarded to the first author.
* Corresponding author. Tel.: +1 519 661 3263; fax: +1 519 661 3495.
E-mail addresses: jhowell@ivey.uwo.ca (J.M. Howell)8 cmshea@cisunix.unh.edu (C.M. Shea)8
chiggins@ivey.uwo.ca (C.A. Higgins).
1
Tel.: +1 603 862 3322; fax: +1 603 862 3383.
2
Tel.: +1 519 661 3269; fax: +1 519 661 3959.

0883-9026/$ - see front matter D 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.jbusvent.2004.06.001
642 J.M. Howell et al. / Journal of Business Venturing 20 (2005) 641–661

and discriminant validity. Preliminary evidence of criterion-related validity indicated that the
champion behavior measure was also positively related to project performance. Together these
analyses provided support for the construct validity of the champion behavior measure.
D 2004 Elsevier Inc. All rights reserved.

Keyword: Champion behavior measure

1. Executive summary

The successful introduction and acceleration of new product development is an important


source of competitive advantage, survival, and renewal for many organizations. Studies of
product innovation success highlight that champions, individuals who informally emerge to
actively and enthusiastically promote innovations through the crucial organizational stages,
are necessary to overcome the social and political pressures imposed by an organization and
convert them to its advantage. While studies emphasize the importance of champions for
keeping innovations alive and thriving, empirical investigations have conceived of champions
as either present or absent in the innovation process. To date, the behaviors of champions, and
the extent to which these behaviors must be enacted to contribute to successful project
performance, have not been specified. This study adds to the innovation and champion
literatures by defining, developing, and validating a measure of champion behavior that
identifies and quantifies what champions actually do to promote innovations successfully in
organizations.
The development of the champion behavior measure involved collection of data from
different samples in multiple stages. First, using the act frequency method, a comprehensive
set of 102 items reflective of champion behavior was developed based on a survey of middle
managers and executives. Second, to determine which of these behaviors constituted the core
of the champion behavior construct, a different managerial sample rated the extent to which
each of the 102 behaviors was representative of champion behavior. Twenty-nine behaviors
were deemed to represent the core of the domain of championship. Finally, we evaluated the
psychometric properties of the champion behavior measure. To examine the factor structure,
reliability, and convergent, discriminant, and criterion-related validity of the champion
behavior measure, 47 product innovations in 13 firms were studied. A total of 47 nominated
champions completed personality measures (i.e., achievement orientation, innovativeness,
and risk-taking propensity), 47 top division managers rated project performance, and 216
innovation team members rated champion behavior and leadership style.
Results from a principal components analysis followed by a confirmatory factor analysis
indicated that of the 29 champion behaviors, 14 met the selection criteria for retention. The
champion behavior measure was composed of three factors:

(1) expressing enthusiasm and confidence about the success of the innovation,
(2) persisting under adversity, and
(3) getting the right people involved.
J.M. Howell et al. / Journal of Business Venturing 20 (2005) 641–661 643

Analyses of the psychometric properties of this 14-item measure indicated acceptable


reliability as well as convergent and discriminant validity. Preliminary evidence of criterion-
related validity indicated that the champion behavior measure was also positively related to
successful project performance. Collectively, these analyses provided support for the
construct validity of the champion behavior measure.
The present study adds to the innovation and champion literatures by developing a
measurement scale of champion behavior with acceptable psychometric properties. We hope
that the development of this measure will spur future empirical studies on champions of
innovation, an under-researched area of inquiry. Furthermore, since the champion behavior
measure developed in the current study reflects the behaviors of later-stage champions of
successful product innovations, future investigations are needed to determine whether these
behaviors apply to early-stage champions, or to those who fail in their innovation promotion
efforts, as well to different types of innovations.
A practical implication of our study is that championship may be a constellation of
behaviors that can be nurtured and learned, a very important fact in light of the significant
linkage between frequency of champion behavior and successful project performance.
Specifically, the measure may be useful for identifying potential champions and for designing
learning experiences to develop the skills of aspiring champions. Providing potential
champions with training and development of these behaviors could help them promote
successful innovations in organizations.

2. Introduction

Introducing new products to the market is an important way by which organizations


adapt or respond to increasing global competition, rapidly changing customer demands,
technological advancements, and shorter product life cycles (Brown and Eisenhardt, 1995).
Yet, despite the enormous creative and innovative effort invested in the generation of new
product ideas, the vast majority of ideas never advance beyond the idea generator’s desktop
(Stevens and Burley, 1997). It is therefore not surprising that there has been a flurry of
theoretical and empirical work on how to organize effectively for innovation in order to
ensure the idea survives until a market test can determine its retention or divestment
(Brown and Eisenhardt, 1995; Dougherty and Hardy, 1996; Kessler and Chakrabarti, 1996;
Venkataraman et al., 1992). Case studies of product innovation success highlight internal
championing as a critical means by which social and political pressures are applied to
overcome organizational inertia and move new product ideas through the phases of small to
large project endeavors, market launch, and ultimate market success (e.g., Achilladelis et
al., 1971; Schon, 1963; Tushman and Nadler, 1986). Moreover, champions who initiate
new product development can provide the organization with opportunities to learn about
new environmental niches, to create new competencies, and to develop strategic options
(Burgelman, 1991). In the words of Frost and Egri (1991: 270), bWithout dedicated
champions, ideas for product innovations may remain dormant for future development and
implementation.Q
644 J.M. Howell et al. / Journal of Business Venturing 20 (2005) 641–661

Champions are defined as individuals who informally emerge in an organization


(Chakrabarti, 1974; Roberts and Fusfeld, 1988; Schon, 1963) and make ba decisive
contribution to the innovation by actively and enthusiastically promoting its progress through
the critical [organizational] stagesQ (Achilladelis et al., 1971: 14). The value of understanding
championship is threefold.3
First, at the levels of firms and business units, champions can significantly affect, both
positively and negatively, the way resources and power are distributed internally, as well as
major strategic actions and performance over time. Second, at the levels of individuals, teams,
and business units, championship can be an important determinant of internal organizational
consequences, such as the speed and distance of career progression, and the motivation and
retention of key talent who may be tempted to leave and become entrepreneurs if their
projects are not receiving sufficient internal sponsorship. Third, champions can foster
substantive, cross-functional communication within firms and stimulate vigorous debate
between managers that can facilitate effective decisions about initial and continuing
investment in innovation projects (Van de Ven et al., 1999).
Although increasing evidence points to the importance of champions for keeping product
innovation ideas alive and thriving (Achilladelis et al., 1971; Burgelman, 1983; Day, 1994;
Dougherty and Hardy, 1996; Frost and Egri, 1991), empirical investigation of these
individuals is limited. Most of what has been reported about champions is largely banecdotal,
reflecting the researcher’s impressions, rather than reliable and valid measurement using well-
accepted instrumentsQ (Howell and Higgins, 1990: 317). In the few instances when
champions have been reliably identified and studied, a significant limitation is the
operationalization of champion emergence as the presence or absence of a champion
(Howell and Higgins, 1990). This operationalization does not acknowledge that there may be
degrees of championship. Furthermore, the behaviors of champions that contribute to
innovation performance have not yet been specified, nor has the extent to which these
behaviors must be enacted to influence project performance.
The current research attempts to contribute to the innovation and champion literatures in
four ways. First, we conceptualize and measure championship, not as a binary phenomenon,
but as a continuous variable. Second, using the act frequency method (Buss and Craik, 1980,
1983, 1985), we identify a comprehensive set of champion behaviors, i.e., what champions
actually do to promote innovation in organizations. Third, using multisource data from
champions, divisional executives, and innovation team members, the psychometric character-
istics of the champion behavior measure are assessed. Finally, implications for using the
champion behavior measure to identify, select, and develop potential champions of
innovations are discussed.
To our knowledge, to date researchers have not attempted to develop a measure of
champion behavior, although there are some instruments that assess perceptions of
champions in general. For instance, in an attempt to identify behaviors that managers might
adopt to spearhead corporate entrepreneurial initiatives, Pearce et al. (1997) developed a

3
We appreciate the insights from an anonymous reviewer who offered a broad and balanced view of championship.
J.M. Howell et al. / Journal of Business Venturing 20 (2005) 641–661 645

theoretically derived measure that distinguished managers who exhibited an entrepreneurial


orientation. However, many of their scale items appear to reflect effective leadership
practices targeted at direct reports (e.g., communicating a strategic vision, building a
cohesive team, showing enthusiasm for skill acquisition), rather than the wide ranging
actions required to promote new ideas within the firm. Whether the managers who
demonstrated entrepreneurial behavior, as seen through the eyes of their direct reports,
actually promoted innovation and, in turn, performance, remain open questions. An
instrument developed by Shane et al. (1995: 938) based on the breceived wisdom of
innovation championingQ, focused on preferences for innovation championing strategies
rather than the actual behavior of champions. Given the limitations of current research,
coupled with the well-documented finding that champions are necessary in order for
innovation and internal corporate venturing to occur (e.g., Burgelman, 1983; Frost and Egri,
1991; Howell and Higgins, 1990; Shane, 1994), it appears that development of a measure of
champion behavior is a worthwhile endeavor.
Drawing on the extant literature, we begin by defining the champion behavior construct.
We then use the act frequency method (Buss and Craik, 1980, 1983, 1985) to develop a
comprehensive set of champion behaviors and determine which of these behaviors constitute
the core of the championship construct. Finally, we construct a measure of championship
based on those core behaviors and assess its psychometric properties.

2.1. The champion behavior construct

In a classic article on radical military innovations, Schon (1963) identified the role of a
champion as pivotal to the successful implementation of an innovation. He observed that
champions buse any and every means of informal sales and pressureQ (1963: 84) and bdisplay
persistence and courage of heroic qualityQ (1963: 85) in promoting an innovation. Maidique
(1980) elaborated on Schon’s observations noting that champions needed to demonstrate
personal commitment to the idea, promote the idea with conviction, persistence, and energy
through informal networks, and willingly risk their position and reputation to ensure success.
Consistent with Schon and Maidique’s conceptualization of champion behaviors, Roberts
and Fusfeld’s (1988) discussion of the roles in the innovation process and Markham et al.’s
(1991) study of research and development projects acknowledged that champions
aggressively sell new ideas to others and put themselves on the line for their cause. In
their research on the internal corporate venturing process, Garud and Van de Ven (1992)
reported that champions persisted in a course of action despite experiencing negative
outcomes.
Further insights into champion behavior were offered by Burgelman (1983). He reported
that champions articulated a convincing master strategy for the idea, and initially mobilized
resources covertly by bact[ing] as scavengers, reaching for hidden or forgotten resources to
demonstrate feasibilityQ (1983: 238) and then more overtly, by establishing and maintaining
contact with top management to keep them informed and enthusiastic about the project.
Similarly, Venkataraman et al. (1992) observed that a new venture idea required a
champion to exert social and political effort to galvanize internal support for the concept
646 J.M. Howell et al. / Journal of Business Venturing 20 (2005) 641–661

and settle conflicts between key stakeholders. Other scholars have noted that champions
engage in coalition building and other cooperative influence tactics in order to develop
solidarity around the idea (e.g., Galbraith, 1982; Markham and Griffin, 1998; Shane,
1994).
Chakrabarti and Hauschild’s (1989) model of the division of labor in innovation
management expanded the conceptualization of champion behaviors beyond selling the idea
and obtaining stakeholders’ support. According to them, championing also involved steering
the goal formation process by evaluating ideas in terms of their fit with organization strategy,
explaining, teaching, and motivating others to get involved, and dealing with opponents by
managing through the firm’s power centers. Howell and Higgins’s (1990) study of the
personality characteristics, influence tactics, and leadership behaviors of champions provided
some support for Chakrabarti and Hauschild’s (1989) model. They reported that champions
could be distinguished from nonchampions by communicating a clear vision of what the
innovation could be or do, displaying enthusiasm about innovation, demonstrating commit-
ment to it, and involving others in supporting it.
Current conceptualizations and empirical work point to several behaviors that appear to
characterize champions. Namely, champions create and communicate strategic meaning
around the innovation, persistently promote the innovation, sell the idea to top management
in order to secure resources, and involve and motivate others to support the innovation. While
the extant literature serves as a useful starting point in pinpointing some champion behaviors,
further research is required to identify a comprehensive set of champion behaviors, to assess
which of these behaviors are core to the domain of championship, and to develop a
continuously scaled measure of these behaviors. In the next section, we describe the method
used to develop a reliable and valid measure of champion behavior.

3. Method

The development and validation of the champion behavior measure involved collection of
data from different samples in three stages. In the first two stages, the act frequency approach
to scale development was used (Buss and Craik, 1980, 1983, 1985). Specifically, the act
frequency method is a technique for identifying behavioral acts at the core, mid-range, and
periphery of dispositions (e.g., dominance; Buss and Craik, 1980, 1983) or organization
behaviors (e.g., self-promotion, voice and loyalty; Cooper et al., 1990).
There are two phases in the act frequency method: act nomination and act prototypicality.
The intent of the first phase is to generate an extensive sample of behavioral acts that fall
within the specified domain, in this case, champions of innovation. The purpose of the second
phase is to identify which of these behavioral acts are central, mid-range, and peripheral to the
specified domain using prototypicality ratings.
Our rationale for selecting the act frequency method to develop the champion behavior
scale was twofold. First, in the act nomination phase, an extensive list of behavioral acts is
generated by respondents that attempt to ensure the content validity of the measure. Given the
limited research on champions of innovation, reliance on the extant literature to construct
J.M. Howell et al. / Journal of Business Venturing 20 (2005) 641–661 647

champion behavior items may artificially restrict the range of behaviors actually displayed by
champions. The use of act nomination helps to ensure that the measure taps a broad spectrum
of behaviors exhibited by champions of innovation. Second, in the prototypicality rating
phase, a different sample of respondents identifies behavioral acts that most represent the core
of the champion behavior construct. Behaviors that are peripheral to the domain of
championship are eliminated. Thus, we believed that the act frequency method was an
appropriate technique to use in order to identify the content and internal structure of the
champion behavior construct.
In the third stage, using a different sample of respondents, the psychometric characteristics
of the champion behavior measure were evaluated including its factor structure, reliability,
and convergent, discriminant, and criterion-related validity. Below we describe the three
stages for developing and validating the champion behavior measure.

3.1. Phase 1: act nomination

The first phase of the study was aimed at generating a fairly exhaustive list of acts
within the domain of championship. Managers at the middle and executive levels who
represented a variety of functional areas were targeted as a sample since it was likely that
they would be associated with innovations and be most familiar with the behavior of
champions. One hundred North American subscribers of a business journal that publishes
articles relevant to practicing executives were contacted by telephone by a research
assistant.
To compile a list of champion behaviors, respondents were asked if they could think of an
innovation that had been implemented within the past year or was currently being
implemented in their organization. As recommended by Buss and Craik (1980, 1983,
1985), the term innovation was defined broadly so that the champion behaviors generated by
respondents were applicable to a wide range of innovations. Specifically, an innovation was
defined as a technology, a product, or a process that was new to the organization. Six people
could not think of a recent innovation, leaving 94 useable responses. Respondents reported a
wide array of innovations and varied levels of success and impact. Examples included
management process innovations such as the implementation of statistical process control and
manufacturing resource planning systems, new manufacturing equipment innovations, and
new product development such as a soft drink, breath freshening gum, and a satellite system.
They were then asked whether there was an individual who stood out in their minds as the
prime supporter and facilitator of the implementation of the innovation. Seventy-three
respondents indicated that there was one such individual. In the remaining cases, the
innovation was viewed as a team effort. The 73 respondents who clearly identified a
champion were then asked an open-ended question: bPlease list five actions which the
individual actually performed that you consider being representative of his or her support of
the innovation.Q A total of 304 acts were nominated.
The study’s investigators independently edited the list of 304 champion behaviors
according to five criteria specified by Buss and Craik (1983): redundancies, non-act
statements, general tendency statements, statements considered too vague to constitute an
648 J.M. Howell et al. / Journal of Business Venturing 20 (2005) 641–661

observable behavior, and statements considered too specific for use in a scale intended for
a range of innovations. The investigators then met to resolve any inconsistencies in their
editing. Editing of the 304 nominated championship behaviors resulted in the following
deletions: 8 items were not behaviors at all (e.g., bbelieved in the innovationQ, bis familiar
with work in the areaQ), 6 items were general tendency statements (e.g., bwas aggressiveQ),
22 items were too vague (e.g., bmanaged the projectQ, binvestigated the changeQ), and 112
items were considered too specific for use in a scale intended for a wide range of
applications (e.g., bstarted monthly progress reviews across the countryQ, bdebugged the
equipmentQ, bissued several memos on the subjectQ). Finally, 54 items closely duplicated
other items and were deleted for redundancy. The final list of champion acts contained
102 items.

3.2. Phase 2: act prototypicality

The second phase of scale development evaluated which of the 102 behaviors generated
during phase 1 were considered to be at the core of the domain of championship using a
different sample of respondents. A questionnaire was developed and mailed to 1000 graduate
business school alumni of a publicly funded university. The school has 16,711 alumni, of
which 79% are male, and are 52 years old on average. Since the targets were individuals who
would be at the middle to upper levels of the organizational hierarchy, only those who had
graduated at least 6 years earlier were included in the sampling frame. A total of 194
questionnaires were returned, i.e., 19.4% response rate. The innovations represented a wide
spectrum and included both successes and failures.
Using the same definition of innovation from phase 1, respondents were asked if they
could think of an innovation that had been implemented within the past year or was
currently being implemented in their organization. Total of 45 of the 194 respondents could
not think of any, reducing the number of useable responses to 149. Respondents were asked
to rate, on a 7-point scale, the extent to which the behaviors listed were representative of
champion behavior. The definition of a champion was the same as phase 1. A b7Q indicated
that the behavior was very representative; a b1Q indicated that it was not representative; and a
b4Q meant that the respondent felt that the behavior fit moderately well. Mean item
prototypicality ratings ranged from 3.08 to 6.46 on a scale from 1 to 7. As specified by the act
frequency method (Buss and Craik, 1983), the acts were divided into quartiles based on their
mean prototypicality ratings, and the 29 acts in the top quartile were considered to represent
the core of the domain of championship (i.e., the cut-off point for the top quartile was 5.87;
see Table 1).

3.3. Phase 3: psychometric characteristics of the champion behavior measure

The purpose of phase three was to evaluate the psychometric properties of the champion
behavior measure. In the following sections, we describe the innovations, sample, procedure,
measures, and data analysis methods used to examine the reliability and validity of the
champion behavior measure.
J.M. Howell et al. / Journal of Business Venturing 20 (2005) 641–661 649

Table 1
Champion behavior items, prototypicality means, and standard deviations
Items Prototypicality Standard
means deviations
Shows personal commitment to the innovation 6.46 0.85
Expresses confidence in what the innovation can do 6.34 0.84
Secures the top level support required 6.25 1.02
Expresses strong conviction about the innovation 6.24 0.85
Keeps pushing enthusiastically 6.24 1.01
Backs the people involved 6.21 0.92
Sells the need to top management 6.21 0.92
Does not give up when others say it cannot be done 6.20 0.99
Sticks with it 6.18 0.94
Stands behind the innovation and supports it 6.18 0.85
Recognizes the potential of the innovation 6.17 0.96
Sells the innovation to key people 6.17 0.86
Understands the relationship between business needs and the innovation 6.12 1.14
Persists in the face of adversity 6.08 1.00
Gets the right people involved 6.08 1.06
Gives ongoing support to the project team 6.08 0.93
Sees the need for the innovation 6.06 1.00
Gets problems into the hands of those who can solve them 6.05 1.18
Shows tenacity in overcoming obstacles 6.03 1.01
Points out reasons why the innovation will succeed 6.00 1.19
Creates awareness of the need for the innovation 5.97 1.22
Shows optimism about the success of the innovation 5.94 1.14
Knocks down barriers to the innovation 5.92 1.19
Gets key decision-makers involved 5.91 1.27
Makes improvements based on feedback 5.90 1.04
Enthusiastically promotes the innovation’s advantages 5.89 0.96
Continues to be involved with the innovation until it is implemented 5.88 1.31
Keeps stakeholders involved in the process 5.88 0.95
Convinces the people involved that it is a good change 5.87 1.02

3.3.1. Description of innovations


In contrast to the first two phases where a broad range of innovations were included in the
generation of champion behavior items and the assessment of item prototypicality, in the
third phase, only product innovations were sampled. Our rationale for assessing the
reliability and validity of the champion behavior measure for product innovations
exclusively was twofold. First, prior research has emphasized the importance of clearly
specifying the type of innovation to be sampled in order to overcome empirical instability of
results arising from research on innovation in complex organizations (Bigoness and
Perreault, 1981; Downs and Mohr, 1976). Second, based on his meta-analysis of
relationships between organization innovation and 13 of its possible determinants,
Damanpour (1991) concluded that the type of innovation and organization should be
distinguished, particularly where the focus is on understanding the generation, development,
650 J.M. Howell et al. / Journal of Business Venturing 20 (2005) 641–661

and implementation of innovations rather than on understanding organizational determinants


such as centralization, specialization, and so on.
A product innovation was defined as a new technology or a combination of technologies
introduced commercially to meet either an external user’s or a market’s need (Kessler and
Chakrabarti, 1996). Four criteria were used to select the innovations: (a) the innovation was
developed internally and was new to the company on the dimensions of products, markets,
and/or technologies; (b) the product had a development cycle of one and one-half to three
years; (c) the innovation required significant investments of company resources to accomplish
a result beyond the year in which the expenditure was made; and (d) the innovation was still
in progress or very recently completed.

3.3.2. Sample
The domain of interest consisted of manufacturing firms involved in new product
development. Our sample included 47 companies in the electronics (Standard Industry
Codes 334 and 335) and automotive and aerospace (Standard Industry Code 336)
industries that had new product innovations in progress or were very recently
implemented. These companies were identified through a survey of 250 Chief Executive
Officers of the largest North American firms in these industries in terms of sales. Based on
the responses, 64 innovations in 22 companies appeared to meet the four selection criteria
for inclusion in the study. Preliminary interviews were conducted with senior executives to
explain the general purpose of the study, to obtain an in-depth description of the
innovation, to identify the product champion(s), and to seek their agreement to participate
in the study.
Specifically, initially via an open-ended question, company executives were asked to
identify a product innovation and the key people involved in the project. To reliably identify
champions, the interviewer read aloud descriptions of four different roles in the innovation
process (i.e., product champion, technical innovator, executive sponsor, and chief executive)
derived from Achilladelis et al. (1971) and asked the executives to identify the person or
persons who clearly fit each role. To minimize attributional bias in identifying champions, the
specific identification of the role as project champion, technical innovator, and so on was not
provided. If no one fits the champion role, the interview ended.
Based on these interviews, 47 product innovations were identified that clearly met the
established criteria. In addition, each of these innovations had a clearly identified product
champion; there were no instances of multiple champions for a project. Thirteen
organizations agreed to participate in the study with an average of four projects each. Each
organization was part of a multinational with sales ranging from $2.3 billion to $125 billion
and the number of employees ranging from 12,000 to 300,000.
The product innovations all involved a significant investment of resources and were new to
the company in terms of products, markets, and technologies. For example, one firm in the
electronics industry developed a remote data storage facility that would allow customers to
keep their data processing applications on-line 24 h a day and quickly recover them in the
event of a disaster. During the development phase, 10 software engineers and 4 hardware
engineers worked full time on the project at an estimated cost of $1.5 million. A second
J.M. Howell et al. / Journal of Business Venturing 20 (2005) 641–661 651

example involved a firm in the automotive industry that established a project team tasked
with overcoming problems associated with spring tension clamps fluctuating due to thermal
variations. The team developed a constant tension band clamp that provides leak-proof joints
for hoses in automotive applications. The project involved 10 people and $3.5 million in
equipment, material, and prototyping labor. Finally, a firm in the aerospace and defense
industry developed a shipboard integrated communication system designed to allow tactical
decision-makers single-terminal access to various incoming communications. This project
was 3 years in the development phase and involved a nine-member team and a total
investment of about $4 million.

3.3.3. Procedure
Interviews were conducted with the 47 identified champions using a structured protocol.
On average, the interviews took 1.5 h. During the interview, the champion described the
product innovation in detail and verified the identity of key individuals involved in the
innovation including the product team members and top division managers.
To minimize same-source bias, three different questionnaires were administered: to
champions to assess their personal characteristics (N=47), to members of the innovation team
to measure leadership style and champion behavior (N=237), and to top division managers to
measure project performance (N=47). Questionnaires were returned to the researchers directly
via external mail. Several follow-up letters and phone calls were made to ensure a high
response rate. All of the questionnaires distributed to champions and top division managers
were returned. The response rate for team members was 91% (N=216). The number of team
member questionnaires returned per champion ranged from two to nine with an average of
five. The demographic characteristics of the sample used to assess the psychometric
characteristics of the champion behavior measure (i.e., champions, division managers, and
innovation team members) are described in Table 2.

3.3.4. Analysis of psychometric characteristics


The psychometric properties of the champion behavior measure were assessed in five
stages: (1) factor structure, (2) reliability, (3) convergent validity, (4) discriminant validity,
and (5) criterion-related validity. Each stage is described below.

3.3.4.1. Factor structure. The factor structure of the champion behavior measure was
assessed using the 29 core champion behaviors generated by the act frequency method (see
Table 1). Team members rated the frequency of observed champion behaviors using a 5-point
scale ranging from not at all (0) to frequently, if not always (4). The multidimensionality of
this construct was evaluated using principal components extraction with varimax rotation.
Using the criterion of eigenvalues greater than one, the factors extracted from the principal
components analysis were assessed with a confirmatory factor analysis to evaluate overall
model fit.

3.3.4.2. Reliability. The internal consistency or reliability of the champion behavior scale
was evaluated using Cronbach’s alpha.
652 J.M. Howell et al. / Journal of Business Venturing 20 (2005) 641–661

Table 2
Demographic characteristics of the sample
Respondent Age Gender Functional area Organization Job tenure Organization
(M) Male Female tenure (M) (M) level
(%) (%)
Champion 44 95 5 Engineering or R&D (44%) 11 years 3 years Middle
General Management (28%) Management
Marketing or Sales (7%)
Production (2%)
Other (19%)
Division 48 100 0 General Management (56%) 13 years 4 years Executive
Managers Engineering or R&D (19%)
Production (10%)
Marketing or Sales (2%)
Other (13%)
Team 41 88 12 Engineering or R&D (53%) 10 years 4 years Lower
Members Marketing or Sales (14%) Management
General Management (8%)
Finance, Accounting or
Information Systems (5%)
Production (4%)
Other (16%)

3.3.4.3. Convergent validity. To determine that the champion behavior measure assessed
aspects of individuals’ promotion of innovation, it was correlated with an established measure
of leadership behavior. Prior research has demonstrated that transformational leadership is
empirically linked to the promotion of innovations in organizations (Howell and Higgins,
1990). Specifically, to introduce innovations successfully in organizations, champions engage
in transformational leadership behaviors including expressing a compelling vision of the
innovation’s potential (charisma), encouraging others to think about creative ideas or novel
solutions to challenges (intellectual stimulation), and developing innovation team members’
capabilities (individualized consideration). Thus, to assess convergent validity, we examined
the correlations between the champion behavior measure and the three components of
transformational leadership (i.e., charisma, intellectual stimulation, and individualized
consideration). Adequate convergent validity would be demonstrated if the champion behavior
measure was moderately correlated with the three transformational leadership subconstructs.
Transformational leadership behavior was measured using 28 items drawn from Bass and
Avolio’s (1990) Multifactor Leadership Questionnaire (MLQ Long Version). Innovation team
members judged how frequently the designated champion engaged in these transformational
behaviors using a 5-point frequency scale (not at all (0), once in a while (1), sometimes (2),
fairly often (3), frequently, if not always (4)). To reduce the 28 items to a manageable number
of factors, a principal components analysis using a varimax rotation was conducted, and the
three expected subconstructs of transformational leadership, namely charisma (6 items,
a=0.91), intellectual stimulation (4 items, a=0.85), and individualized consideration (5 items,
J.M. Howell et al. / Journal of Business Venturing 20 (2005) 641–661 653

a=0.90), were identified. Items were summed to create a scale score for each of the three
transformational leadership subconstructs.

3.3.4.4. Discriminant validity. Three measures were used to establish discriminant validity
by demonstrating that the champion behavior measure does not simply reflect the personality
characteristics of champions. Previous research has reported that champions manifest the
personality characteristics of risk-taking propensity, innovativeness, and achievement
orientation (Howell and Higgins, 1990). If indeed the champion behavior measure reflects
perceptions of the focal champion’s enthusiastic promotion of innovation rather than his/her
personality characteristics, then champion behavior scale scores should be relatively
uncorrelated with risk-taking propensity, innovativeness, and achievement orientation.
Champions completed two scales from the Jackson Personality Inventory (Jackson, 1984)
that measured risk-taking propensity and innovativeness, and one scale from the Personality
Research Form E (Jackson, 1987) that measured achievement orientation. A true/false
response format was used for these measures. The number of items, Cronbach’s alpha, and
conceptual definition for each scale were: risk-taking propensity (21 items, a=0.69): enjoys
taking chances; innovativeness (19 items, a=71): develops novel solutions to problems; and
achievement orientation (17 items, a=0.45): aspires to accomplish difficult tasks. With the
exception of the achievement orientation measure, these scale reliabilities were considered
acceptable based on Nunnally’s (1978) criterion of 0.70 or greater. Given the low reliability of
the achievement orientation measure, it was dropped from further analyses.
To assess discriminant validity, for each champion, mean scores were calculated for each
of the three subconstructs of champion behavior based on innovation team members’ ratings.
These mean scores were then correlated with champions’ self-assessment of their personality
characteristics.

3.3.4.5. Criterion-related validity. Criterion-related validity concerns the relationship


between the champion behavior measure and other variables with which the measure is
supposed to correlate. Many studies have shown that the presence of a champion is positively
correlated with product innovation success (e.g., Burgelman, 1983, 1991; Day, 1994; Kessler
and Chakrabarti, 1999; Markham and Griffin, 1998). Thus, if the champion behavior measure
predicts project performance, it will be an indication of criterion-related validity.
We evaluated criterion-related validity by testing the relationship between the champion
measure and project performance. Top division managers evaluated the project’s performance
using a measure developed by Keller (1986). Five items were used to rate the project’s
technical quality, budget and cost performance, meeting assigned schedules, value to the
organization, and overall project performance on a 5-point Likert scale (1=very low; 5=very
high). The five items were summed to create an average score for the project performance
measure. The Cronbach’s alpha for the scale was 0.86.
One potential problem with the project performance measure was restriction of range. Very
few projects were rated low (i.e., most projects were successful). Obviously, in studies
involving champions, it is hard to uncover unsuccessful projects that individuals are willing
to discuss.
654 J.M. Howell et al. / Journal of Business Venturing 20 (2005) 641–661

To assess criterion-related validity, a regression was conducted using the three


subcomponents of the champion behavior measure as independent variables and project
performance as the dependent variable.

4. Results

4.1. Factor structure

The 29 core champion behaviors were analyzed to assess the multidimensionality of the
construct using principal components extraction with varimax rotation. Three factors met the
criterion of eigenvalues greater than one and explained 64.40% of the variance in champion
behavior. Factor (1) explained 54.20% of the variance, factor (2) explained 5.40%, and factor
(3) explained 4.80% of the overall variance in the 29 items. Items were retained in a factor if
they had a loading at or above 0.65 on that factor, and cross-loadings at or below 0.40 on the
other two factors. Table 3 shows the retained and dropped items together with the loadings
and cross-loadings.
Factor 1 contained six items relating to the champion’s enthusiasm and confidence about
the success of the innovation. Factor 2 was comprised of six items focusing on the
champion’s persistence under adversity. Factor 3 consisted of three items relating to the
champion’s efforts at getting the right people involved in the innovation.
The three-factor structure obtained from the principal components analysis was subjected
to a confirmatory factor analysis. The chi-square statistic was highly significant (v 2
(87)=165.20, pb.001). The root-mean-square residual was 0.026 (below 0.03 is considered
acceptable), and the root-mean-square error of approximation was 0.065 (below 0.10 is
considered acceptable). The goodness-of-fit index was 0.903 (above 0.90 is considered
acceptable) while the adjusted goodness-of-fit index was 0.87 (above 0.90 is considered
acceptable). Overall, the model showed adequate fit (Bentler and Bonett, 1980). However,
examination of the modification indices indicated that one item (bDoes not give up when
others say it cannot be doneQ) from the persistence under adversity factor was problematic
(i.e., its measurement error was highly correlated with measurement error from other items).
This item was dropped and a second confirmatory factor analysis was conducted. The results
showed an improved model fit with a large decrease in the chi-square value (v 2 (74)=116.90).
Furthermore, the root-mean-square residual decreased to 0.024 (from 0.026), and the root-
mean-square error of approximation dropped to 0.052 from 0.065. The goodness-of-fit index
improved to 0.93 (from 0.903) and the adjusted goodness of fit index was 0.90 compared to
0.87. The t-values of each estimated parameter were significant at pb.001. Overall, these
indices suggest a reasonable model fit (Bentler and Bonett, 1980).

4.2. Reliability

The overall Cronbach’s alpha coefficient for the champion behavior scale was 0.94,
indicating strong internal consistency. In addition, the internal consistency reliabilities for the
J.M. Howell et al. / Journal of Business Venturing 20 (2005) 641–661 655

Table 3
Champion behavior item loadings and cross-loadings
Component
1 2 3
Retained Items
Enthusiastically promotes the innovation’s advantages 0.75 0.26 0.29
Expresses strong conviction about the innovation 0.74 0.40 0.13
Expresses confidence in what the innovation can do 0.73 0.18 0.23
Shows optimism about the success of the innovation 0.69 0.38 0.25
Points out reasons why the innovation will succeed 0.66 0.26 0.37
Keeps pushing enthusiastically 0.66 0.40 0.31
Sticks with it 0.36 0.75 0.23
Shows tenacity in overcoming obstacles 0.19 0.72 0.32
Continues to be involved with the innovation until it is implemented 0.34 0.71 0.34
Knocks down barriers to the innovation 0.29 0.69 0.34
Does not give up when others say it cannot be done 0.39 0.68 0.11
Persists in the face of adversity 0.28 0.65 0.29
Gets problems into the hands of those who can solve them 0.06 0.32 0.76
Gets the right people involved 0.24 0.30 0.73
Gets key decision makers involved 0.31 0.27 0.69

Dropped Items
Sees the need for the innovation 0.64 0.35 0.15
Shows personal commitment to the innovation 0.48 0.55 0.29
Backs the people involved 0.26 0.54 0.52
Gives ongoing support to the project team 0.26 0.51 0.39
Secures the top level support required 0.41 0.17 0.65
Makes improvements based on feedback 0.22 0.45 0.62
Sells the need to top management 0.53 0.13 0.60
Sells the innovation to key people 0.58 0.17 0.58
Keeps stakeholders involved in the process 0.27 0.48 0.57
Convinces the people involved that it is a good change 0.49 0.27 0.55
Understands the relationship between business needs and the innovation 0.42 0.30 0.49
Eigenvalues 15.70 1.60 1.40
Percent (%) of variance explained 54.20 5.40 4.80

three subconstructs exceeded Nunnally’s criterion of 0.7: the Cronbach’s alphas were 0.91 for
expressing enthusiasm and confidence about the success of the innovation, 0.90 for persisting
under adversity, and 0.83 for getting the right people involved (see Table 4).

4.3. Convergent validity

Table 4 shows the correlations between the three subconstructs of the champion behavior
measure and the three subconstructs of transformational leadership. The correlations ranged
from a low of 0.41 (between charisma and getting the right people involved) to a high of 0.73
(between charisma and expressing enthusiasm and confidence about the success of the
656 J.M. Howell et al. / Journal of Business Venturing 20 (2005) 641–661

Table 4
Reliabilities, convergent validity, and discriminant validity of the champion behavior measure
Champion behavior Cronbach’s Convergent validity (N=216) Discriminant validity (N=216)
subconstructs alpha Charisma Intellectual Individualized Risk taking Innovativeness
(N=216) stimulation consideration
Expresses 0.91 0.73** 0.53** 0.49** 0.11 0.11
enthusiasm
and confidence
Persists under 0.90 0.56** 0.61** 0.57** 0.05 0.11
adversity
Gets the right 0.83 0.41** 0.57** 0.58** 0.03 0.14
people involved
** pb.01.

innovation). The average correlation between the subconstructs was 0.56. These moderate to
high correlations indicate that the champion behavior scale has adequate convergent validity.

4.4. Discriminant validity

Table 4 shows the correlations between the three subconstructs of the champion measure
and the two personality characteristics, namely, risk-taking propensity and innovativeness.
The correlations are low, ranging from 0.03 to 0.14. These results suggest that the champion
behavior measure is not merely a reflection of the champion’s personality characteristics, and
provide strong support for the discriminant validity of the champion behavior measure.

4.5. Criterion-related validity

The regression relating the three subconstructs of the champion behavior measure to
performance was significant ( F(3,43)=3.91, pb.05) and explained 21.9% of the variance in
performance. This finding provides a preliminary indication of criterion-related validity.

5. Discussion

The present study attempted to advance the extant literature by defining, developing, and
validating a scale for measuring champion behavior. Using the act frequency method, we
first generated a list of 102 different behaviors that middle- to upper-level managers
identified as representative of the behaviors of a champion involved in a recent innovation
in their organization. We then obtained a prototypicality ranking of the 102 behaviors from
a second sample of managers. This ranking enabled us to identify 29 behaviors that can be
considered to reside at the core of the domain of championship. As might be expected, our
study indicated that the 29 items representing the top quartile of championship acts were
not unidimensional. Rather, a principal components analysis followed by a confirmatory
factor analysis yielded a three-factor solution containing 14 items. These analyses indicated
J.M. Howell et al. / Journal of Business Venturing 20 (2005) 641–661 657

that the champion behavior construct is comprised of three separate, yet interrelated
components.
The results indicated that the 14-item champion behavior measure demonstrated acceptable
psychometric properties including high internal consistency reliabilities and face validity. The
measure’s three factors, labeled as expressing enthusiasm and confidence about the success of
the innovation, getting the right people involved, and persisting under adversity, were largely
consistent with previous literature on champions. One factor, demonstrating enthusiasm and
confidence in the innovation, was consistent with Howell and Higgins (1990) who found that
champions could be distinguished from non-champions by their display of high expectations
for, and optimism about, the innovation, and Burgelman’s (1991) observations of the
enthusiastic zeal of new product development champions at Intel. The factor concerning
getting the right people involved, including key decision makers, has been highlighted as a
crucial part of the championing role by many researchers who have emphasized that
champions need to obtain support for the innovation and convince top management that the
project is worthwhile (e.g., Burgelman, 1983; Chakrabarti and Hauschild, 1989; Markham et
al., 1991; Roberts and Fusfeld, 1988; Shane, 1994). The factor persisting under adversity was
consistent with the writings of all of the above authors who invariably stress the staying
power of champions.
The assessment of the convergent validity of the champion behavior scale indicated
moderate to high correlations with transformational leadership. This is not surprising as
previous studies (Howell and Higgins, 1990) have shown that champions are informal
transformational leaders. Nonetheless, the behaviors represented here go beyond those assessed
using a measure of transformational leadership and provide a more accurate assessment of what
champions of innovation actually do to promote innovations in organizations. The champion
behavior scale also showed high levels of discriminant validity with the personality
characteristics of innovativeness and risk-taking propensity. Thus, the assessment of champion
behavior can be clearly distinguished from champions’ personality characteristics.
Testing the criterion-related validity of the champion behavior measure represented an
initial attempt at establishing the construct validity of this measure. Previous research has
demonstrated that the presence of a champion was positively related to higher levels of new
product development performance that, in turn, positively affected firm level performance
(Markham et al., 1991; Markham and Griffin, 1998). Using a continuous measure of
championship, the present study supported the established positive relationship between
champion behavior and project performance. However, it is important to note that there was
restriction of range on the performance measure, so the findings here need to be tempered
given this limitation.

5.1. Limitations

One limitation of the present study was the use of cross-sectional data. Accordingly, causal
interpretations of the results were precluded. A second limitation was that a subjective
measure of performance was used to assess criterion-related validity. With regard to the
statistical analyses, the sample size of 47 champions was a third limitation of the present
658 J.M. Howell et al. / Journal of Business Venturing 20 (2005) 641–661

study. However, it is important to note that only a few organizational members ever choose to
become champions; it is a relatively rare organizational phenomenon (Howell and Higgins,
1990; Shane, 1994). Moreover, in order to eliminate self-report bias, it is important to identify
champions through a rigorous nomination process, such as the one used in the present study
to validate the measure. The combination of the rare occurrence of champions, and the need
to reliably identify them, implies that undertaking research on champions requires
considerable investment of time and resources.
A fourth limitation was that, at different phases in the development of the champion
behavior measure, such as the act prototypicality phase, there were low response rates.
However, this limitation is compensated for, in part, the 100% response rates from
champions and division managers, and responses from multiple innovation team members
(an average of five team members for each champion) in evaluating the psychometric
properties of the measure. Thus, despite the limitation of some low response rates, we do
have a fairly thorough knowledge of this sample of champions from different and multiple
respondents.
Fifth, the study’s sample was predominantly male; therefore, it is unknown whether the
relationships reported here generalize to female champions. Sixth, while the first two phases
of this study contained a broad range of innovations, the validation phase involved only new
product innovations in manufacturing organizations.
A final limitation is that the champion behavior measure developed in the current study
represents individual behaviors associated with product innovations that had already achieved
a relatively high level of internal organizational success. Failed innovations or early-stage
innovations were not sampled in the present study. It is conceivable that early-stage or
unsuccessful champions might exhibit a wider or narrower range of behaviors, and possibly
behaviors that differ from the behaviors assessed by the measure developed here. Moreover,
early-stage or unsuccessful champions may not enjoy the visibility or positive internal
reputation of champions associated with successful innovations. Senior executives may not
know who they are, and their colleagues may view their behaviors in a negative (e.g.,
disruptive and annoying), rather than a positive (e.g., proactive and engaged) light. Thus, this
measure reflects behaviors of later-stage champions whose innovations were still in progress
or recently completed.

5.2. Future research

The shortcomings of the current study provide future research opportunities. First, the
measure developed in the current study needs to be tested with another sample of champions,
in particular, champions associated with early-stage innovations and with unsuccessful
innovations. There is also a need to validate the champion behavior scale with other types of
innovations and in different organizations. Another worthwhile research endeavor is to
investigate how the behaviors of later-stage champions differ from bstopping-championsQ of
failing projects (Royer, 2001). While there are commonalities between these types of
champions with respect to their personal characteristics and influence tactics, stopping-
champions differ from champions on several dimensions including using various socio-
J.M. Howell et al. / Journal of Business Venturing 20 (2005) 641–661 659

political tactics and their networks to rally supporters to halt the failing project, posing
strategic and financial arguments for project withdrawal, and involving team members in
establishing decision criteria to evaluate whether to continue or exit the project. Based on
Royer’s (2001) analysis, it is conceivable that some unique behaviors are associated with the
stopping-champion roles that are not captured in the current champion behavior measure.
Second, increasing the size and gender diversity of the sample, and gathering longitudinal
data on champion behavior and innovation performance using an objective measure would be
worthwhile directions for future research. Finally, there is a need for cross-cultural studies of
champion behavior, especially in light of Shane et al.’s (1995) findings that preferences for
championing strategies vary depending on national culture.
The scale developed in the current study provides a vehicle for future research to shed light
on the process and impact of champions in organizations. Markham et al. (1991) have pointed
out that there has been conflicting views in the innovation literature about the degree to which
an individual must engage in champion activities before he or she is considered to be
championing an innovation. In other words, what is the threshold that must be reached before
an individual is perceived as a full-fledge champion? Rather than being restricted to viewing
championship as an ball or nothingQ phenomenon, this scale enables researchers to measure
later-stage champion behavior as a continuous variable.
As Venkataraman et al. (1992) have noted, very little research has been done on how the
process of championing takes place within firms, what mechanisms are used by champions to
foster new ventures, and which mechanisms are successful or not. Even in the decade
following the publication of the article of Venkataraman et al., there remains little research on
how champions actually promote new ventures in organizations. Furthermore, how
champions trigger the action thresholds of senior management to pay attention to new ideas,
needs, and opportunities (Van de Ven, 1986) is a largely unexplored research question even
today. Future investigations could examine how later-stage champions select, package, and
sell product innovation ideas to enhance their personal credibility and to prompt management
to attend to these ideas. Moreover, building on the work of Van de Ven et al. (1999), the
balance and timing of the pluralistic leadership roles in the innovation process, including
champions, institutional leaders, critics and mentors, and the impact on ultimate project
success represents a fruitful area for future inquiry.

5.3. Implications

A practical implication of our study is that championship may be a constellation of


behaviors which can be nurtured and learned, a very important fact in view of the significant
linkage between frequency of champion behavior and project performance. Using behaviors
as indicators of championship extends beyond the question of bwho are champions?Q to ask
bwhat do they do?Q Specifically, the measure may be useful for identifying and selecting
potential champions, and for designing learning experiences to develop their skills. For
instance, individuals who aspire to be champions could learn how to present ideas in a
compelling way, to use multiple influence tactics skillfully and strategically, and to enlist
support and overcome resistance of key stakeholders.
660 J.M. Howell et al. / Journal of Business Venturing 20 (2005) 641–661

While champions have a substantial impact on decreasing cycle time and are associated
with more products reaching the market (Kessler and Chakrabarti, 1999; Markham and
Griffin, 1998), they do not always produce winning products or assure market or financial
success of a new venture (Markham et al., 1991). The measure developed here highlights that
later-stage champions keep pushing enthusiastically for the innovation and persist in the face
of obstacles, and suggests that senior managers may be swept away by champions’ passion
and conviction and potentially ignore danger signs that the project is failing. This implies that
executives need to be vigilant in their active support of champions and use explicit and
rigorous criteria to assess the project’s progress.
Research has shown that, in the electronics, automotive, and aerospace industries, it takes
as many as 3000 raw ideas to yield one successful new product in the marketplace (Stevens
and Burley, 1997). New product development projects proceed from the raw idea through
various stages of approval for additional support and resource allocation in organizations.
Champion behaviors include those that are instrumental in successfully guiding projects
through the approval hurdles: displaying persistence, expressing strong conviction in the
innovation, and involving key individuals. Selecting individuals who display these particular
champion behaviors may increase the probability that projects make it through the approval
process.
To ensure that novel ideas are brought to life, expanding our knowledge of the origins and
contexts of champions is vital. The current study constitutes a first attempt to develop and
validate a measure of later-stage champion behavior. We hope that the combination of high
face validity and sound psychometric characteristics will encourage continued use and
development of the champion behavior scale.

References

Achilladelis, B., Jervis, P., Robertson, A., 1971. A Study of Success and Failure in Industrial Innovation.
University of Sussex Press, Sussex, England.
Bass, B.M., Avolio, B.J., 1990. Manual for the Multifactor Leadership Questionnaire. Consulting Psychologist
Press, Palo Alto, CA.
Bentler, P.M., Bonett, D.G., 1980. Significance tests and goodness-of-fit in the analysis of covariance structures.
Psychol. Bull. 88, 588 – 606.
Bigoness, W.J., Perreault, W.D., 1981. A conceptual paradigm and approach for the study of innovators. Acad.
Manage. J. 24, 68 – 82.
Brown, S.L., Eisenhardt, K.M., 1995. Product development: past research, present findings and future directions.
Acad. Manage. Rev. 20, 343 – 378.
Burgelman, R.A., 1983. A process model of internal corporate venturing in the diversified major firm. Adm. Sci.
Q. 28, 223 – 244.
Burgelman, R.A., 1991. Intraorganizational ecology of strategy making and organizational adaptation: theory and
field research. Organ. Sci. 2, 239 – 262.
Buss, D.M., Craik, K.H., 1980. The frequency concept of disposition: dominance and prototypically dominant
acts. J. Personal. 48, 379 – 392.
Buss, D.M., Craik, K.H., 1983. The act frequency approach to personality. Psychol. Rev. 90, 105 – 126.
Buss, D.M., Craik, K.H., 1985. Why not measure that trait? Alternative criteria for identifying important dis-
positions. J. Pers. Soc. Psychol. 48, 934 – 946.
J.M. Howell et al. / Journal of Business Venturing 20 (2005) 641–661 661

Chakrabarti, A.K., 1974. The role of champion in product innovation. Calif. Manage. Rev. 17, 58 – 62.
Chakrabarti, A.K., Hauschild, J., 1989. The division of labor in innovation management. R&D Manage. 19,
161 – 171.
Cooper, W.H., Dyke, L., Kay, P., 1990. Developing act frequency measures of organizational behaviors. Academy
of Management Best Paper Proceedings, San Francisco.
Damanpour, F., 1991. Organizational innovation: a meta-analysis of determinants and moderators. Acad. Manage.
J. 34, 555 – 590.
Day, D.L., 1994. Raising radicals: different processes for championing innovative corporate ventures. Organ. Sci.
5, 148 – 172.
Dougherty, D., Hardy, C., 1996. Sustained product innovation in large, mature organizations: overcoming
innovation-to-organization problems. Acad. Manage. J. 39, 1120 – 1153.
Downs, G.W., Mohr, L.B., 1976. Conceptual issues in the study of innovation. Adm. Sci. Q. 21, 700 – 714.
Frost, P.J., Egri, C.P., 1991. The political process of innovation. In: Staw, B.M., Cummings, L.L. (Eds.), Research
in Organizational Behavior. JAI Press, Greenwich, CT, pp. 229 – 295.
Galbraith, J.R., 1982. Designing the innovating organization. Organ. Dyn. 10, 5 – 25.
Garud, R., Van de Ven, A.H., 1992. An empirical evaluation of the internal corporate venturing process. Strateg.
Manage. J. 13, 93 – 109.
Howell, J.M., Higgins, C.A., 1990. Champions of technological innovation. Adm. Sci. Q. 35, 317 – 341.
Jackson, D.N., 1984. Jackson Personality Inventory. Sigma Assessment Systems, Port Huron, MI.
Jackson, D.N., 1987. Personality Research Form E. Sigma Assessment Systems, Port Huron, MI.
Keller, R.T., 1986. Predictors of performance of project groups in R&D organizations. Acad. Manage. J. 29,
715 – 726.
Kessler, E.H., Chakrabarti, A.K., 1996. Innovation speed: a conceptual model of context, antecedents, and
outcomes. Acad. Manage. Rev. 21, 1143 – 1191.
Kessler, E.H., Chakrabarti, A.K., 1999. Speeding up the pace of new product development. J. Prod. Innov. Manag.
16, 231 – 247.
Maidique, M.A., 1980. Entrepreneurs, champions and technological innovation. Sloan Manage. Rev., 59 – 76.
(Winter).
Markham, S.K., Griffin, A., 1998. The breakfast of champions: associations between champions and product
development environments, practices and performance. J. Prod. Innov. Manag. 15, 436 – 454.
Markham, S.K., Green, S.G., Basu, R., 1991. Champions and antagonists: relationships with R&D project
characteristics and management. J. Eng. Technol. Manag. 8, 217 – 242.
Nunnally, J.C., 1978. Psychometric Theory. McGraw-Hill, New York.
Pearce II, J.A., Kramer, T.R., Robbins, D.K., 1997. Effects of managers’ entrepreneurial behavior on subordinates.
J. Bus. Venturing 12, 147 – 160.
Roberts, E.B., Fusfeld, A.R., 1988. Critical functions: needed roles in the innovation process. In: Katz, R. (Ed.),
Managing Professionals in Innovative Organizations. Harper Collins, New York, pp. 101 – 120.
Royer, I. 2001. Stopping-champions of Failing Projects. Paper presented at the Academy of Management
Conference, Washington, D.C. August.
Schon, D.A., 1963. Champions for radical new inventions. Harvard Bus. Rev. 41, 77 – 86.
Shane, S.A., 1994. Are champions different from non-champions? J. Bus. Venturing 9, 397 – 421.
Shane, S., Venkataraman, S., MacMillan, I., 1995. Cultural differences in innovation championing strategies. J.
Manage. 21, 931 – 952.
Stevens, G.A., Burley, J., 1997. 3000 raw ideas=1 commercial success! Res. Technol. Manag. 40, 16 – 28.
Tushman, M.L., Nadler, D., 1986. Organizing for innovation. Calif. Manage. Rev. 28, 74 – 92.
Van de Ven, A.H., 1986. Central problems in the management of innovation. Manage. Sci. 32, 590 – 607.
Van de Ven, A.H., Polley, D.E., Garud, R., Venkataraman, S., 1999. The Innovation Journey. Oxford Univ.
Press, New York.
Venkataraman, S., MacMillan, I., McGrath, R., 1992. Progress in research on corporate venturing. In: Sexton, D.
(Ed.), State of the Art in Entrepreneurship. Kent Publishing, New York.

View publication stats

You might also like