Vous êtes sur la page 1sur 11

Administrative Science Quarterly March 1989 v34 n1 p110(22) Page 1

Entrepreneurs in academe: an exploration of behaviors among life


scientists.
by Karen Seashore Louis, David Blumenthal, Michael E. Gluck and Michael A. Soto

This paper explores entrepreneurship in the research university, a setting in which there has been
a marked change over the past half century in norms governing relationships between scholars
and the commercial sector. A survey of life science faculty members in research universities is
used to distinguish five types of academic entrepreneurship: (1) engaging in large-scale science
(externally funded research), (2) earning supplemental income, (3) gaining industry support for
university research, (4) obtaining patents or generating trade secrets, and (5) commercialization -
forming or holding equity in private companies based on a faculty member’s own research. The
results suggest models for the different types of entrepreneurship. Individual characteristics and
attitudes are the most important predictors of large-scale science and supplemental income,
which are more traditional, while local group norms play a more important role in predicting active
involvement in commercialization. University policies and structures have little effect on
entrepreneurship. Implications for organizational theory and the role of the university are
discussed. (Reprinted by permission of the publisher.)

© COPYRIGHT 1989 Cornell University, Johnson there is a critical relationship between entrepreneurship
Graduate School and the overall competitiveness of our corporate sector in
the world economy. The role of entrepreneurship in
Entrepreneurs in Academe: An Exploration of Behaviors revitalizing ossified or traditional organizations is of
among Life Scientists particular social importance (Peterson, 1981) and is
considered to be one form of strategic management
This paper explores entrepreneurship in the research (Mintzberg, 1973). Snow and Hrebiniak’s (1980) analysis
university, a setting in which there has been a marked suggested that entrepreneurial organizational strategies,
change over the past half century in norms governing those based on rapid commercialization of new inventions,
relationships between scholars and the commercial sector. engender higher performance in industries operating in
A survey of life science faculty members in research uncertain environments. Interest in the relationship of
universities is used to distinguish five types of academic entrepreneurship to organizational performance and vitality
entrepreneurship: (1) engaging in large-scale science is not limited to the private sector but is also being widely
(externally funded research), (2) earning supplemental discussed in educational contexts (e.g., Etzkowitz, 1983;
income, (3) gaining industry support for university Mazzoni, 1987). In this paper we report a study that
research, (4) obtaining patents or generating trade secrets, examined academic entrepreneurship, defined as the
and (5) commercialization--forming or holding equity in attempt to increase individual or institutional profit,
private companies based on a faculty member’s own influence, or prestige through the development and
research. The results suggest models for the different marketing of research ideas or research-based products.
types of entrepreneurship. Individual characteristics and Universities are not usually viewed as leaders in
attitudes are the most important predictors of large-scale entrepreneurship. In fact, there is often a tendency to
science and supplemental income, which are more distinguish between the search for truth in science,
traditional, while local group norms play a more important considered a legitimate function of the university, and the
role in predicting active involvement in commercialization. search for invention, which is considered an inappropriate
University policies and structures have little effect on focus on ideas that have potential commercial or practical
entrepreneurship. Implications for organizational theory applicability (Ravetz, 1971; Wade, 1984). Nevertheless,
and the role of the university are discussed.(*) there has been a notable increase in the number of
scientists and science watchers who champion increased
INTRODUCTION There is increasing consensus in the entrepreneurship in universities. Entrepreneurship is
organizational and management literature that believed to contribute to the rapid movement of scientific
entrepreneurship is a significant factor in organizational ideas into the commercial arena (Blumenthal, Gluck, and
effectiveness. According to Benveniste (1987), risk taking Louis, 1985), to provide a critical contribution of scientists
and accepting responsibility are interdependent and to the national economy and society (Ping, 1980), to
equally important to an effective professional organization. revitalize the scientific endeavor through new sources of
Peters and Waterman (1982) linked entrepreneurship with research funds, and to contribute to the university’s
invention and innovation and argued that it is causally financial base through royalties on patents (Blumenthal et
related to productivity, while Kanter (1983) suggested that al., 1986b).

- Reprinted with permission. Additional copying is prohibited. - GALE GROUP


Information Integrity
Administrative Science Quarterly March 1989 v34 n1 p110(22) Page 2

Entrepreneurs in academe: an exploration of behaviors among life


scientists.
Sources and Distribution of Entrepreneurship (Baldridge and Burnham, 1975; Daft and Becker, 1978;
Rosenblum and Louis, 1981). Along these lines, Kanter
Whether entrepreneurship is considered good or bad, it is (1983) argued that the matrix structure supports
clearly not evenly distributed: some institutions and entrepreneurship. Other writers have emphasized the
individuals demonstrate it more than others. Explanations importance of policies and practices, such as reward
for this variable distribution fall into patterns that are systems, that may stimulate individual or group
familiar to organizational theorists. Organizational entrepreneurship (Kerr and Slocum, 1987). Despite the
psychologists tend to emphasize individual characteristics high levels of interest in entrepreneurship, there is
and attitudes (such as achievement motivation) as the remarkably little systematic data on the nature of
source of entrepreneurship (McClelland et al., 1976). entrepreneurship in the university or other nonbusiness
Although achievement motivation does not have strong settings. Discussions about what stimulates university
predictive power (Peterson, 1981), studies of academics faculty to be more entrepreneurial are similarly
indicate that other individual attributes may play a part. For speculative. The research reported below begins to fill that
example, Liebert (1977) indicated that past success (as gap, using data obtained from life scientists located in
measured by research publications) is associated with research-intensive universities. This paper has two main
effective "grantsmanship." Age and gender may also be purposes: (1) to describe five different types of
related to incentives. More established scientists may have entrepreneurship and their incidence and patterns of
more to "sell," may be less motivated by traditional occurrence in the population of research-intensive
academic incentives (tenure, disciplinary awards), which universities and (2) to examine a variety of questions that
they have already achieved, and may have greater are more directly related to organizational theory, including
financial incentives, such as having children in college (a) the structure of entrepreneurship as a behavioral
(Zuckerman and Merton, 1972; Etzkowitz, 1983). Women, construct, (b) the relation between individual
who have tended to be less scientifically productive, may entrepreneurship and several classes of possible
also be less likely to be entrepreneurial (Cole and predictors drawn from the literature, (c) whether there is
Zuckerman, 1984). Attitudes can play a part as well. such a phenomenon as an entrepreneurial elite, either at
Etzkowitz (1983) and Peters and Fusfeld (1982) argued the individual or institutional level, and (d) whether there is
that some scientists seek out industry associations any institutional patterning that suggests that universities,
because they are predisposed to commercializing their like other organizations, have distinctive entrepreneurial
ideas, in contrast to scientists who stumble across a strategies.
marketable finding or wait for industry to take the initiative
to seek them out. Another line of speculation concerns the METHOD The analysis presented below is based on two
importance of cultural support for entrepreneurship. surveys, both conducted in 1985, one of a sample of life
Research indicates that local culture is more important in scientists located in major research universities and the
this respect than broad social values (Peterson, other of key administrators in the same universities. First,
1981:70-71), a point strongly supported by Kanter (1983: university administrators were surveyed in the 50
129-138) and consistent with classic organizational institutions that receive the most federal research funds.
investigations such as Seashore’s (1954). In the world of Data on university policies and characteristics were
organized science, Pelz and Andrews (1976) noted, collected in a telephone survey of the 40 university
colleagues in the work group have an impact on the administrators who had been identified as having the most
behavior of individual scientists. This local contextual responsibility for the life science departments included in
effect is not related to the size of the work group (Cohen, the study. The interviews were conducted by trained
1981) but to the tendency for members to conform to local professional interviewers. When necessary, other
norms of behavior regarding entrepreneurship (Peters and university administrators were also contacted to obtain
Fusfeld, 1982). Local behavioral norms can be reinforced complete information. For the faculty survey, only
over time through recruiting, socialization, and retention institutions where the key administrator had responded
(Van Maanen, 1976). A recent analysis of relationships were included, which eliminated 10 of the top 50
between life scientists and industry found dense institutions. For those 40 universities, the 3,180 life
institutional networks that are interpreted as an effect of science faculty members listed in published catalogs as
local norms about entrepreneurship (Ennis, 1986). A final members of the departments of biochemistry, molecular
factor that may account for entrepreneurship is the biology, genetics, microbiology, biology, cellular biology, or
organizational structures and policies that may affect such botany were identified (Peterson’s Guides, 1984). From
activities. Previous research suggests that the size, this list 1,594 individuals were randomly selected. Faculty
complexity, and authority structure of the organization will members in the sample were mailed an eight-page
be associated with innovativeness in educational settings questionnaire dealing primarily with his or her research
- Reprinted with permission. Additional copying is prohibited. - GALE GROUP
Information Integrity
Administrative Science Quarterly March 1989 v34 n1 p110(22) Page 3

Entrepreneurs in academe: an exploration of behaviors among life


scientists.
activities. If the questionnaire was not returned within three dollars per year. Supplemental income. After World War II,
weeks, they received a second mailing and a follow-up the belief spread rapidly that scientists could maintain the
telephone call. One hundred fifty-six respondents were ideal of basic research without sacrificing contact with the
ineligible (deceased, retired, no longer associated with the world of practice (Etzkowitz, 1983). Most universities
university, or incorrectly reported as a faculty member in explicitly condone limited consulting, and some form of
the catalog). Of the remaining eligible respondents, 69 income augmentation is the norm for most academics. The
percent (997) completed questionnaires. Missing data at typical amount of money earned from selling personal
the item level in the two surveys reduced the number of scientific expertise is not great, however, and the impacts
usable faculty responses to 778 from 40 universities. of consulting on scholarly performance are limited (Boyer
and Lewis, 1985). Supplemental income, which explicitly
Academic Entrepreneurship excludes unearned income, was measured by asking what
percentage over basic salary the faculty member earned in
In this paper we define five basic forms of academic recent years. A seven-point categorical scale was used, on
entrepreneurship: (1) large-scale science (obtaining large, which 1 = none, 2 = 1-10 percent, 3 = 11-20 percent, 4 =
externally funded research projects), (2) earning 21-30 percent, 5 = 31-40 percent, 6 = 41-50 percent, and 7
supplemental income outside the university, mainly = over 50 percent. The median response was 2.26, with a
through consulting (knowledge transfer for personal gain), mode of 2. About a fifth of the scientists had no
(3) soliciting funds from industry (capitalizing on supplemental income at all, while half earned no more
university-industry relationships to provide new sources of than 10 percent over their base salary. Fewer than 5
funding for research), (4) patenting the results of research, percent supplemented their income by 40 percent or more,
and (5) forming companies based on the results of a figure that is somewhat lower than for a 1975 random
research. Although all forms of academic entrepreneurship sample of full-time university faculty in all disciplines
stimulate occasional controversy in the academic (Marsh and Dillon, 1980). Not all forms of supplemental
community, the types are roughly ordered from the most to income are viewed as entrepreneurial, however, i.e.,
the least compatible with a traditional view of the teaching additional courses in the summer. Respondents
university-based scientist’s role (Etzkowitz, 1983; Krimsky, were asked to check the top two sources of supplemental
1984; Wade, 1984). income from a list of 10. Most of the life scientists’ extra
income was derived from activities that might be
Large-scale science. Academic science increasingly considered modestly entrepreneurial, involving the sale of
requires big laboratories and a large staff. This has the individual’s expertise through nonuniversity
affected the basis for evaluating individual performance: employment (10 percent), consulting for profit-making (27
the size and number of research grants has come to be an percent) and nonprofit (18 percent) firms, and the lecture
indicator of the individual’s disciplinary competence and circuit (19 percent). The least common major sources of
prestige (Liebert, 1977). University budget processes have income were the most entrepreneurial (compensated
been shown to reinforce the importance of grantsmanship directorships and royalties from licenses, with only 1
(Pfeffer and Salancik, 1974; Pfeffer and Moore, 1980). percent each). The most traditional forms of earning
This type of entrepreneurship may be most challenging for supplemental income--teaching extra courses and
younger scholars, who try to establish their reputations by royalties from books--provided significant sources of
developing laboratories of their own (Merton, 1968). income for only 5 and 7 percent, respectively. An estimate
Individual involvement in large-scale science was of actual supplemental income was calculated using the
measured by the amount of externally funded research for response to this question and that from another question
which the respondent was the principal investigator during concerning the respondent’s salary. Salary categories
the 1984-85 academic year. The median for the sample is were as follows: less than $20,000, $20,000-29,999,
$195,000 per year, exclusive of overhead. This level of $30,000-39,999, $40,000-49,999, $50,000-59,999,
funding is sufficient to fund a modest laboratory, with a $60,000-69,999, and $70,000 or more. Consulting income
small staff of semiprofessional technicians and perhaps a was estimated by multiplying the midpoint value of the
few doctoral students. However, the standard deviation is respondent’s income category by the midpoint value of his
rather high ($285,000), suggesting that there is or her consulting category. The mean salary for the
considerable variation in this type of entrepreneurship population was $50,775, with a standard deviation of
even within a sample of faculty members associated with $14,997. The median estimated supplemental income was
research-intensive universities. Ten percent of the faculty a modest $4,843. In 1975, the average for all faculty was
members get $3,000 or less of external funding per year, approximately $2,700, which in 1984 dollars would be
while the top 20 percent of faculty get $251,000 or more of $5,415. The standard deviation is quite high, however
external funding, with a few receiving several million ($7,198). For the nearly 17 percent of the respondents
- Reprinted with permission. Additional copying is prohibited. - GALE GROUP
Information Integrity
Administrative Science Quarterly March 1989 v34 n1 p110(22) Page 4

Entrepreneurs in academe: an exploration of behaviors among life


scientists.
who augmented their income by more than 20 percent, the universities: 19 percent had applied for or been granted a
estimated average supplemental earnings were nearly patent or had generated a trade secret based on their
$27,000. Industrial support for university research. research. Approximately one-third of the respondents
Recently, the organization of industrial research and basic indicated that research support from industry or consulting
science research in the university has become increasingly to industry contributed significantly to the work on which
similar (Peters and Fusfeld, 1982; Blumenthal et al., patents were based. Direct commercial involvement.
1986a, 1986b). This has led to exchange of personnel, Etzkowitz (1983) pointed out that the emerging
common research projects and, in some cases, characteristics of large-scale science provide faculty
large-scale joint ventures. There are a variety of members with the management skills that permit easier
motivations for scientists to seek funding from industry, but entry into the private sector.
scientists who obtain money from this source are more
likely to select research problems because of their The formation of private firms whose products are based
potential commercial applicability (Blumenthal et al., on the university scientists’ own research is a logical
1986b). This supports the contention that this form of extension of the trends listed above. This form of
entrepreneurship is less traditional than the two previously entrepreneurship is the most non-traditional and
discussed, although it may be more easily available. In controversial in that it involves potential use of university
addition, more than 50 percent of our respondents facilities and graduate students to meet the firm’s
indicated that research support from industry "provides commercial goals (Blumenthal et al., 1986b). Direct
resources for research that could not be obtained commercial involvement was measured by asking
elsewhere" and "involves less red tape than federal respondents whether they held equity in companies whose
funding." Finally, the market for obtaining industry support products and services were based on their own research.
is less tied to the applicant’s past productivity than This form of entrepreneurship is the least common: only
federally funded research, which may make industry more about 7 percent indicated that they held equity in such
attractive to younger scholars or others with weak track companies, and only a handful held equity in more than
records (Liebert, 1977). Industry funding was measured by one.
calculating the proportion of the total of the respondent’s
externally funded grants and contracts budget that came The Structure of Entrepreneurship
from private industry. The median research support from
industry was 7.7 percent; again, however, the standard The literature has not addressed the key question of
deviation is rather high (21 percent), which demonstrates whether there is an emerging group of "entrepreneurial
wide variability in this regard. Only 23 percent of life scholars" who engage in multiple forms of
science faculty members receive some funding from entrepreneurship. If the different forms of entrepreneurship
industry, suggesting that this behavior is still the exception identified above cluster empirically, then such an academic
rather than the norm. Of those who do receive such type may be emerging. On the other hand, if the
funding, fewer than half get more than 25 percent of their associations between different types of entrepreneurship
external research funding from industry; the mean is 34 are not high, then we are drawn to conclude that the above
percent (see also Blumenthal et al., 1986b). However, characteristics may be a consequence of different
there is a small proportion of faculty members who might motivations, impulses, or opportunities and represent very
be assumed to be industry dominated: about 7 percent different styles of adaptation to the changing scientific and
receive more than half of their external research funds scholarly environment. Table 1 presents a correlation
from industry. Patenting. Patenting is a logical extension of matrix showing the relationships among the variables.
the tendency toward increasing interest in commercially Although the associations are statistically significant, they
applicable results. The incidence of patents awarded to are not particularly large. The Pearson correlation between
university scientists or universities has been growing, and equity holding and supplementary income is .33, and this
many universities now have patent offices or have is the strongest relationship in the table. Thus, we reach
stimulated independent foundations to deal with patents the preliminary conclusion that the "entrepreneurial
and royalties (Blumenthal et al., 1986b). In addition, many scholar" is not common among life scientists. This pattern
biotechnology companies report that they have made of relationships does not provide a strong justification for
patent applications based on research that they have creating a summary scale of entrepreneurship, and a
funded in universities (Blumenthal, Gluck, and Louis, principal components analysis (not shown) produced only
1985). Patent involvement was measured by whether the a weak first factor. Thus, in the remainder of this analysis
respondent had applied for or been granted patents or had we treat entrepreneurship as a multidimensional concept.
generated trade secrets. Patenting behavior still involves a
minority of life science faculty members in major research Characteristics Predicting Entrepreneurship
- Reprinted with permission. Additional copying is prohibited. - GALE GROUP
Information Integrity
Administrative Science Quarterly March 1989 v34 n1 p110(22) Page 5

Entrepreneurs in academe: an exploration of behaviors among life


scientists.
Individual-level variables. Measures of individual calculating the local norms variable for the five
demographic, career, and attitudinal characteristics in our entrepreneurial behaviors was as follows: [Mathematical
survey included: type of appointment, which measured Expression Omitted] where k = the university, [X,sub,ik] =
whether the individual was located in a medical school; the measure for individual i in university k; and [n,sub,k] =
professional age, measured as years since completing the the number of individuals in university k. The individual i
doctoral degree; gender, a dummy variable measured as 1 was included in the calculations of [X,sub,k] for record i,k.
= male; risks to science, measured by a battery of Each institution had between 20 and 45 responses, so the
questions indicating the degree to which involvement with use of this simple model may increase the correlation
industry represents a potential risk to traditional scientific between the X measured at the individual level and at the
values. Faculty members were asked to indicate whether contextual level by a maximum of 5 percent. This was
each of the following posed a great risk, some risk, only a considered tolerable for the exploratory analysis presented
little or no risk: creating pressure for faculty members to in this paper. The local norm measures do not reflect work
spend too much time on commercial activities, shifting too groups that have routine face-to-face interactions, because
much emphasis toward applied research, undermining most major research universities in our sample had
intellectual exchange and cooperative activities within several life science departments. Rather, they are
departments, creating conflict between faculty members contextual peer groups, as defined by location and role. A
who support and those who oppose such activities, substantial body of research supports the utility of using
altering the standards for promotion and tenure, reducing contextual effects of this type in studies of individual
the supply of talented university teachers, and creating behavior (see Burstein, 1980, for a methodological
unreasonable delays in the publication of new findings. review). The group-level measures and their indicators,
Responses were added to form the summary risk scale. along with the correlations between the individual and local
Two other individual-level variables were measured: norm variables were as follows: industry funding, or the
attitudes about university-industry relations, a single item mean percentage of research funding from industry within
indicating whether respondents would like to see the the life science departments in the university (r = .25);
involvement between their university and industry increase consulting patterns, measured as the mean proportion of
a lot, somewhat, stay about the same, decrease, or income over and above base salary earned from
decrease a lot; and professional productivity (the number consulting (r = .29); funded research, the mean amount of
of articles published during an average three-year period external research funding for life scientists at the university
over the respondent’s professional lifetime). Logarithmic (r = .35); patenting, the percentage of faculty members that
transformations were made of the following variables, have a history of patenting (r = .26); equity holding, the
which were not normally distributed: size of research
budget, percent of research budget from industry, percentage of life scientists that hold equity in a firm that
supplemental income, and number of publications. Another uses their research (r = .28); and productivity, the mean
set of individual characteristics that may contribute to a number of articles published over the past three years by
prediction of any particular form of entrepreneurship is, of life scientists at the university. Organizational level
course, the individual’s other entrepreneurial behaviors variables. Organizational structures and policies
(OEBs), e.g., all entrepreneurial behaviors other than that supporting entrepreneurship vary widely. Some
being predicted. These were shown to be modestly universities have large and complex support units (patent
intercorrelated and are therefore likely to predict each offices) and create institutional incentives (seed money
other. In this analysis OEBs were initially treated as a grants to support faculty members’ search for external
separate group of predictors, in order to further investigate funding) (Peters and Fusfeld, 1982). Auspices may also be
the structure of entrepreneurship behaviors. Group-level important: In general, state universities are viewed as less
variables. Local norms, defined as the way in which most supportive of entrepreneurship than private universities.
members of the organization behave, are also likely to However, some land grant colleges and schools with a
influence behavior. For example, a faculty member located technical focus have strong traditional ties with industry,
in a university where many other faculty members engage while others have been encouraging patenting for some
in heavy consulting with private industry may be more time (Peters and Fusfeld, 1982). Universities can also
likely to do the same than one located in a university encourage or discourage faculty consulting and
where such consulting is uncommon. Measures of local involvement in commercialization through the development
norms supporting entrepreneurship were developed by and enforcement of policies (Wade, 1984). Administrative
calculating the mean of the responses of the life scientists support data were obtained from the survey of university
within each university in the sample for each of the five administrators. In each case the administrator’s response
entrepreneurship variables and attaching the mean of the was linked for analysis to the individual faculty member’s
university to the file of the individual. The formula for file. The measures were as follows: auspices, whether the
- Reprinted with permission. Additional copying is prohibited. - GALE GROUP
Information Integrity
Administrative Science Quarterly March 1989 v34 n1 p110(22) Page 6

Entrepreneurs in academe: an exploration of behaviors among life


scientists.
institution is a state or private university; seed money to examine them separately. Table 3 shows the results of
support, whether or not money is provided to support the four regression models that were computed. The first
faculty members in writing grant and contract proposals for two entered individual/university variables as a first step
external funding; patent office size, the number of and OEBs or local norms as a second; the third entered
professional employees in the university patent office; individual/university/local norms as a first step and OEBs
traditional industry ties, measured by an administrator’s as a second step; the last entered
response to a single three-point item indicating how strong individual/university/OEBs as a first step and local norms
the university’s relationships with industry have been in the as a second step. Perhaps the clearest finding from this
past; and university entrepreneurship, a summary scale table is that university administrative support has little
indicating the number of mechanisms that the university effect on entrepreneurship. In Table 2, the largest amount
has for commercializing the research of its faculty of variance explained by university administrative support
members. Scale items included holding equity in variables was 3.8 percent. University reputation (over
faculty-owned firms, holding equity in firms employing which university administrators have little short-term
faculty members, holding equity in firms providing support control) is the only institutional characteristic that enters
to faculty members, having a research foundation to invest more than one equation (not tabled), which reinforces the
in faculty firms, having an office or center for stimulating conclusion that university policies have little direct impact
faculty companies, investing venture capital in life-science on faculty entrepreneurial behavior. When OEBs (Table 3,
firms, and donating land or space to science parks and Model 1) or local norms (Model 2) and organizational
other commercial enterprises. Also measured were the structure and policy variables are included in the same
number of life science patents held by the university and equation, no university variables achieve significance (not
reputation on the National Academy of Science’s average shown). Overall, individual characteristics, other
quality rating on a 1-5 scale of all of the sampled entrepreneurial behaviors, and local norms appear to be
departments within each university (Jones, Lindzey, and about equally effective in terms of [R,sup,2] in explaining
Coggeshall, 1982). entrepreneurship, except in the case of size of research
budget, where individual predictors dominate. Models 1
RESULTS and 3 in Table 3 indicate, however, that the association
between OEBs--or the "entrepreneurship profile"
Relative Importance of Predictor Groups presented by the individual respondent--and size of
research budget and funding from industry is sharply
Our initial approach to answering the questions regarding reduced when other variables are entered into the model
the effects of different predictor groups involved looking at as a first step. In contrast, their strong impact on
the relative contribution of each of the four groups of supplemental income and equity holding is maintained.
predictors: individual demographic characteristics and The independent contribution of OEBs to patenting lies in
attitudes, other entrepreneurial behaviors, group-level between. Local norms of behavior significantly increase
variables measuring local norms, and organizational the multiple [R,sup,2] for all of the dependent variables
variables measuring institutional characteristics and between 4 and 5 percent when added as a second step
policies. Several ordinary-least-squares regression models after other variables (Models 2 and 4). Models 3 and 4
were calculated. In the first, five forms of entrepreneurship confirm that OEBs and local norms are measuring different
were regressed on each of the four groups of predictors effects. For all five forms of entrepreneurship, local norm
separately. This analysis examined the relative importance variables are significant predictors even when controlling
of each group considered by itself; the names of those for OEBs, and vice versa. In the case of the less traditional
variables whose t statistics were significant at the .10 level forms of entrepreneurship --industry funding, patenting,
or greater are shown for informational purposes only. and equity holding--individual characteristics are less likely
Because the other entrepreneurial behaviors and the local to be significant predictors when local norm and OEB
norm variables were composed from the same survey variables are added to the regression models. For these
items, it seemed prudent to explore their distinctive forms of entrepreneurship, OEBs alone account for 27, 40,
contributions to entrepreneurship. Burstein (1980) argued and 50 percent, respectively, of the total [R,sup,2] that is
that distinct estimates can be made in regressions in which attained by the full model, and local norm variables
individual and contextual effect measures are included but account for 41, 32, and 23 percent, respectively, of the
not where "frog pond effects" (the difference between the total [R,sup,2] that is attained by the full models.
individual i’s score on X and the contextual score on X) are
also part of the equation. However, given the existing Evaluation of Specific Predictors
controversies about the use of contextual effects data
based on attitudes in organizational research, we decided A second approach to answering the questions posed at
- Reprinted with permission. Additional copying is prohibited. - GALE GROUP
Information Integrity
Administrative Science Quarterly March 1989 v34 n1 p110(22) Page 7

Entrepreneurs in academe: an exploration of behaviors among life


scientists.
the beginning of this paper involved looking at the specific except OEBs, the [R.sup.2] statistics are highest for size of
predictors that best account for each form of research budget. The overall adjusted [R.sup.2] in the full
entrepreneurship. Table 4 shows the results of five regression (Table 4) is .288, which is substantially larger
regression models that were computed using the 16 than in the other equations. Individual characteristics and
individual, OEB, and local norm variables. We excluded attitudes are relatively more important as predictors of size
organizational variables because they were insignificant in of research budget than for the other forms of
previous regressions. Only variables whose regression entrepreneurship, accounting for .22 of the .29 [R.sup.2]
coefficients are significant at the .10 level or better are (78 percent) that is attained by the full model. Local norm
reported in the table. Two individual-level variables were variables are also significant (particularly the size of the
related to several types of entrepreneurship; as shown in research budget of colleagues), but OEBs barely reach
Table 4. First, the individual’s concern about the risks to significance and are less important predictors of this form
science from working closely with industry was negatively of entrepreneurial behavior than other forms. The pattern
associated with three forms: supplemental income, of significant variables suggests that entrepreneurship
industry funding, and patenting. This suggests that the based on grantsmanship may coexist easily with traditional
deeper their concerns about protecting basic science from academic values: It is associated with high levels of
pressures to commercialize, the less likely scientists are to scientific productivity, is not associated with any significant
behave in an entrepreneurial manner. However, the causal contacts with industry (and presumably other sources of
relationship is unclear: Scientists may change their money for applied research), and flourishes in contexts in
attitudes in order to diminish dissonance between their which other scientists are also productive and engaging in
own behavior and their interpretation of the scientific value large-scale scientific endeavors. In fact, we might relabel
system. Alternatively, exposure to entrepreneurship may the research budget variable as elite entrepreneurship.
convince the scientists of the robustness of basic science These elite entrepreneurs are also younger. Supplemental
against corruption through such activities. Second, the income. The pattern for supplemental income is somewhat
individual’s publication rate in refereed journals was similar but with clear differences, as well. First, the
positively associated with three of the five entrepreneurial relationship between this form of entrepreneurship and
behaviors, industry funding and equity holding being the individual characteristics is more limited (individual
exceptions. Thus, it appears that scientists who meet the variables account for 60 percent of the total explained
highest (quantitative) standards of productivity are most variance). Second, the elite entrepreneur pattern is
likely to be entrepreneurial. Table 4 consistently indicates complicated by the addition of new predictors. In particular,
that the local norm variable corresponding to the behavior OEBs are more important, accounting for 28 percent of the
being analyzed matters the most. In other words, the total variance explained. Among the OEBs, only industry
individual’s entrepreneurship of a given type is strongly funding is not a significant predictor of supplemental
predicted by the local norms for the same form of income. A final difference is that those who earn larger
entrepreneurship. Local university publication rate was amounts of extra income are likely to be older and to be
also significant in two cases, research budget and located in arts and sciences departments (as opposed to
supplemental income. This finding is discussed in more medical schools). It is worth speculating about the
detail below. Each of the OEBs has a significant effect on contrasting associations between age and these two more
at least two other forms of entrepreneurship. Two OEBs common forms of entrepreneurship. On the one hand, it
stand out: Both patenting and supplemental income are seems probable that incentives to become involved in
significant predictors of all other forms of entrepreneurship. large, externally funded research projects are greatest
among those who are on the fast track in major
Predicting Forms of Entrepreneurship universities but have not yet necessarily reached the peak
of their scientific recognition. The motivation to compete
Table 4 may be used to address another theoretical issue for grants may decline as the scientist’s position in the
posed above. Rather than asking about the relative impact prestige hierarchy stabilizes. Because older scientists are
of different categories of predictors, we can elaborate on more visible, they are more likely to be sought out as
the question posed earlier about the structure of academic consultants. On the other hand, this may be a cohort
entrepreneurship. If entrepreneurship actually consists of a effect: Younger scientists are more likely to be in
variety of distinctive behaviors, the regressions should two-career families (minimizing the need to supplement
produce different patterns of significant predictors. income), or they may be contributing to new norms about
Research budget. The most traditional form of the appropriate scale of scientific endeavor (Etzkowitz,
entrepreneurship--size of research budget--is by far the 1983). Funding from industry. Funding from industry
easiest form of entrepreneurship to predict. Table 2 presents a somewhat more complicated picture, partly
indicates that for every category of predictor variables because the level of prediction is weaker than for the other
- Reprinted with permission. Additional copying is prohibited. - GALE GROUP
Information Integrity
Administrative Science Quarterly March 1989 v34 n1 p110(22) Page 8

Entrepreneurs in academe: an exploration of behaviors among life


scientists.
models (only 11 percent of the variance is explained by the members from zero to one. To form this variable, faculty
full regression model in Table 4). No individual members were grouped by their university affiliation.
demographic characteristics enter the equations. University groups were then ranked in descending order
Attitudinal variables (the risks-to-science index and the according to the proportion of faculty members in the
variable measuring approval of increasing university who exhibit the behavior. The Y axis measures
university-industry ties) are associated, but with the causal the cumulative proportion of faculty members with that
inference problems noted above. The individual effects of behavior on a scale from zero to one. The diagonal line
OEBs are modest compared to the effects of belonging to represents an equal distribution of the behavior across all
a group of life scientists who get money from industry. It is universities. Discrepancies between the diagonal and the
clear that the model specified here is not a good predictor, curve are a visual representation of the extent to which
and an alternative organizational-level explanation is behavior is institutionally concentrated. These graphs
explored below. Patenting and equity holding. The models reveal that the statistically significant associations
for the most extreme forms of entrepreneurship--patenting correspond to what might be considered socially
and equity holding--are more similar to each other than significant-facts. As might be expected from the discussion
they are to either of the two most traditional forms in terms above, the highest levels of concentration are found for
of the predictive power of the variables and the relative equity holding: 75 percent of all faculty members holding
importance of different predictor groups. Yet, there are still equity in a company whose products or services are based
some differences between them that emerge in Table 4. on their own research are located in universities containing
Equity holding is better predicted (18 percent of the only about 37 percent of the faculty members in the
variance explained) with a simpler model. Only the sample. Similar levels of concentration are found in the
percentage of life scientists in the university who hold case of industrial funding: 70 percent of the faculty
equity and the OEBs (other than size of research budget) members who receive 25 percent or more of their research
exhibit standardized regression coefficients that are budget from industry sponsors are located in institutions
significant at the .05 level or better. This supports the that contain 40 percent of the faculty members. Fifty
contention that equity holding is the most extreme form of percent of all of the faculty members who have research
entrepreneurship: The variables that significantly predict it budgets above $100,000 are located in institutions that
are other nonscholarly entrepreneurial behaviors and contain only 32 percent of the life scientists in the sample;
being in a context in which entrepreneurship is the norm, 50 percent of those whose income is supplemented by
with the former being by far the most important, accounting more than $8,000 a year over their base salary are located
for 50 percent of the explained variance. Patenting, in institutions containing only 35 percent of the sample. We
however, is both more complex and less stable across the also identified universities that scored in the top quartile on
different regression analyses. Like equity holding, the the measures of entrepreneurship. The results are shown
relative importance of local behavioral norms is very in Table 5. We know that the types of entrepreneurship are
important, making up nearly a third of the explained modestly associated at the individual level. If we look at
variance. All OEBs are significant predictors in Table 4. the most entrepreneurial groups (identified by university
affiliation), there are apparent associations at this level as
Institutional Patterns of Entrepreneurship well. For example, among the ten universities that have
the highest proportion of faculty members with research
The final theoretical question posed at the beginning of budgets larger than the sample median, six are also
this paper concerns the degree to which these data reveal among the top quartile on two or three other forms of
any evidence that there are entrepreneurial universities, entrepreneurship. Only one institution is characterized
rather than just isolated entrepreneurial academics. The solely by high levels of external funding. Similarly, of the
above analyses reveal a consistent finding: For each form ten universities that are in the top quartile on percentage of
of entrepreneurship, the aggregated variable measuring faculty members holding equity, six are also in the top
local behavior on this dimension is among the most quartile on supplemental income, while only two are in the
powerful predictors. To what degree does this statistical top quartile of percentage of research funding from
association actually reflect a concentration of faculty with industry. The ranges of faculty behavior among the top
certain types of behavior in particular institutions? To quartile are as follows: mean research budget over
examine concentration we generated graphs, shown in $195,000, 80 percent (University of Washington) to 42
Figure 1, that display the concentration of a given percent (Northwestern); mean supplementary income over
entrepreneurial behavior within universities and the 10 percent, 59 percent (M.I.T.) to 17 percent (Penn State);
association between different forms of entrepreneurship at mean industry research funding over 25 percent, 38
the level of the university as a whole. In Figure 1, the X percent (Oregon State) to 12 percent (Case Western);
axis measures the cumulative proportion of all faculty percent of faculty holding patents, 40 percent (University of
- Reprinted with permission. Additional copying is prohibited. - GALE GROUP
Information Integrity
Administrative Science Quarterly March 1989 v34 n1 p110(22) Page 9

Entrepreneurs in academe: an exploration of behaviors among life


scientists.
Washington) to 12 percent (Baylor); mean percent holding "science watchers." Clark (1983) noted that the evolution
equity, 44 percent (M.I.T.) to 26 percent (Harvard). There of less prestigious higher education institutions is mimetic:
is a significant exception to this generalization: obtaining Where Harvard, Stanford, and M.I.T. lead in
money from industry. Most of the universities that have entrepreneurship, will the others be far behind?
close funding ties with industry are not entrepreneurial on
any other dimension. Furthermore, several of those that Organizational Theory
score highest on this form of entrepreneurship, but not on
any other, are located in public land-grant institutions. The analysis supports a tentative conclusion that, at least
These institutions may be engaged in long-term in academic settings, entrepreneurship is not an either/or
relationships with state-based industries, and the condition, nor are the different forms of entrepreneurship
association may not necessarily be a consequence of a minor variations on a similar social phenomenon. The data
broader local culture of entrepreneurship. suggest that the most distinctive patterns occur in the case
of obtaining large research grants. The pattern of
DISCUSSION AND CONCLUSIONS associations shown throughout the paper suggests that
"elite entrepreneurs" are not likely to be drawn toward
Entrepreneurship and Science more extreme forms of entrepreneurship. However, there
are also unique patterns associated with each of the other
Many scientists still believe that the search for truth is different forms. Under the assumption that academics are
inconsistent with any interest in profiting from ideas. This not unique in their motivations and behaviors, we may infer
view is, however, increasingly controversial as that research on entrepreneurship in private firms might
policymakers turn to science as a vehicle for energizing also benefit from efforts to identify different patterns and
our national economy and society, and administrators and types. Several of the entrepreneurial forms discussed here
faculty members try to secure more money from both may have cognates in other settings. For example, R&D
industry and state and national governments to support entrepreneurs in industry may be quite distinct from those
their research programs. Irrespective of the position taken, who are effective in bringing products to market or in
our data suggest that life scientists in research-intensive organizing new firms. Our cross-sectional data do not
universities are modestly entrepreneurial. However, permit drawing definitive conclusions about the causes of
despite concerns about weakening the basic science academic entrepreneurship. If, however, we look across all
mission of the university (Krimsky, 1984; Wade, 1984; of the data a number of hypotheses may be made. Until
Varrin and Kukich, 1985), there is little evidence in our recently, research (and popular writing) on
survey to suggest that most life scientists are more entrepreneurship has tended to focus on individual
interested in commercial activities than traditional scientific demographic, educational, and employment characteristics
endeavors. Small minorities are involved in more extreme rather than on the characteristics of the organizations in
forms of commercial entrepreneurship, and these forms of which entrepreneurs are located or the groups in which
entrepreneurship are not strongly associated with running they work. This study suggests that individual
a large-scale externally funded research endeavor. In characteristics provide weak and unsystematic predictions
other words, there is no evidence to suggest that a new of the forms of entrepreneurship that are at the center of
kind of "entrepreneurial scholar" has taken over most discussions about the importance of individual action in
universities. The data suggest that scientifically productive revitalizing organizations and the economy. This was
scholars are more entrepreneurial on several dimensions. rather surprising, since several of the individual variables
Thus, this investigation supports Etzkowitz’s (1983) that we examined--gender and age, for example--have
argument that entrepreneurial behavior has evolved been shown to be relatively strong predictors of other
naturally within the scientific community and is not faculty members’ behaviors such as publication rates
incompatible with maintaining the outward manifestations (Cole and Zuckerman, 1984), and our analyses suggest
of scholarship. However, scholarly productivity is not an that they are very important in predicting the more
important predictor of the more commercial forms of traditional forms of academic entrepreneurship. To test the
entrepreneurship, which supports the argument that they robustness of this conclusion, we also conducted
may be less compatible with traditional university values. additional analyses using alternative individual
Our data also suggest that most academic groups do not characteristics (such as rank and actual age) and looked
develop norms that encourage multiple forms of at additional attitudinal batteries in the survey. None of
entrepreneurship: only six institutions appear distinctively these analyses suggested a powerful effect of
entrepreneurial. However, the fact that a number of the individual-level variables. We therefore hypothesize, based
most prestigious universities are entrepreneurial on on our data, that for the more nontraditional forms of
multiple dimensions provides support for the concerns of entrepreneurship, individual characteristics are moderated
- Reprinted with permission. Additional copying is prohibited. - GALE GROUP
Information Integrity
Administrative Science Quarterly March 1989 v34 n1 p110(22) Page 10

Entrepreneurs in academe: an exploration of behaviors among life


scientists.
by institutional location. There are four possible below. This may, of course, be a finding that is relevant
explanations for the relatively strong effect of local norms only to organizations, such as universities, that are
on individual behavior that would account for the data "loosely coupled" (Weick, 1976). However, the ability of a
regarding the concentration of entrepreneurs in particular large organization to maintain a very strong
institutions: (1) self-selection may produce value and entrepreneurial culture without middle-level support and
behavior consensus (individuals are drawn to these reinforcement may be questionable even in more tightly
settings because they are known to be supportive of or structured settings. Overall, since having a productive
tolerate entrepreneurship); (2) behavioral socialization may faculty appears to be so critical, the traditional strategy of
operate within a work group (individuals are affected by continually supporting the recruitment of the best people in
the behavior of their immediate colleagues and tend to act the field is a precondition to the effectiveness of other
like them); (3) organizational culture may be a factor (a policies that may stimulate (or control) entrepreneurship.
broader set of institutional policies, procedures, and values (*)This research was supported by the Andrew Mellon
reinforces attitudes and behavior regarding Foundation and the Department of Health and Human
entrepreneurship); or (4) strategic management may be a Services, grant DHHS-100A-83. We thank our colleagues,
factor (some universities use recruitment to position Thomas A. Louis, Jack Fowler, Stanley E. Seashore,
themselves in the forefront of changing patterns of Ronald G Corwin, James Hearn, and David Wise for their
academic behavior in order to reap the potential benefits in helpful comments on earlier drafts, as well as three
increased prestige and income). We cannot determine anonymous ASQ reviewers. The remaining flaws are, of
which of these is operating, and this issue is worth further course, our own. REFERENCES Baldridge, J. Victor, and
exploration. We suspect, however, that it is likely that all of Robert A. Burnham 1975 "Organizational innovation:
the alternative explanations contribute, in part, to our Individual, organizational and environmental impacts."
findings. Thus, the fact that the university policies and Administrative Science Quarterly, 20: 165-176.
structures identified here have little impact on faculty Beneveniste, Guy 1987 Professionalizing the
entrepreneurship should not be taken to rule out an Organization. San Francisco: Jossey-Bass. Blumental,
institutional effect, although it suggests that institutions David, Michael Gluck, and Karen Seashore Louis 1985
cannot easily engineer entrepreneurship. The measures of "Prospecting in academe." Unpublished paper, Kennedy
university policies were based on actions that university School of Government, Harvard University. Blumenthal,
administrators identified as institutional responses to David, Michael Gluck, Karen Seashore Louis, and David
changing patterns of entrepreneurship, but local norms of Wise 1986a "Industrial support of university research in
behavior (as measured here) are also an institutional biotechnology." Science, 23: 242-246. 1986b
characteristic and may be a consequence of a variety of "University-industry research relationship in
other policies and practices that we have not examined. biotechonology: Implications for the university," Science,
For example, the definition and enforcement of policies 23, 1361-1366. Boyer, Carol, and Darrell Lewis 1985 And
relating to consulting or conflict of interest varies quite on the Seventh Day: Faculty Consulting and Supplemental
widely among research institutions (Louis, Anderson, and Income. Washingtonm DC: Association for the Study of
Swazey, 1988), and this may send significant messages Higher Education. Burstein, Leigh 1980 "The analysis of
about how faculty members are expected to behave. In multilevel data in educational research and evaluation." In
addition, departments rarely have complete autonomy in David C. Berliner (ed.), Review of Educational Research,
defining personnel needs, and this provides another 8: 158-233. Washington, DC: American Educational
leverage point for administrators. We hypothesize that Research Association. Clark, Burton R. 1983 The Higher
these and other policies and procedures reflect underlying Education System. Berkeley: University of California
values and cultural assumptions about what constitutes Press. Cohen, J. E. 1981 "Publication rate as a function of
appropriate entrepreneurial behavior. Given the range of laboratory size in three biomedical research institutions."
controls available to administrators, and the difficulty of Scientometics, 3: 467-487.
managing organizational cultures in large institutions with
unclear and conflicting missions, fostering or controlling Cole, Jonathan, and Harriet Zuckerman 1984 "The
entrepreneurship from the top may be less effective than productivity puzzle: Persistance and change in patterns of
working at the departmental or division level. The publication of men and women scientists." In Advances in
recruitment of key individuals who may help to alter or set Motivation and Achievement, 2: 217-258. Beverly Hills, Ca:
new behavioral norms or the use of task forces to Sage. Daft, Richard L., and Selwyn W. Becker 1978 The
investigate or recommend changes may help to frame new Innovative Organization. New York: Elsevier-North
expectations about behavior. Developing specific policies Holland. Ennis, James G. 1986 "University-corporate links
may send a signal, but the organization is basically very in biotechnology: The network structure of dual affiliations."
dependent on behavioral expectations that are reinforced Unpublished paper, Department of Sociology, Tufts
- Reprinted with permission. Additional copying is prohibited. - GALE GROUP
Information Integrity
Administrative Science Quarterly March 1989 v34 n1 p110(22) Page 11

Entrepreneurs in academe: an exploration of behaviors among life


scientists.
University. Etzkowitz, Henry 1983 "Entrepreneurial Scientific Knowledge and Its Social Problems. Oxford:
scientists and entrepreneurial universities in American Clarendon Press. Rosenblum, Sheila, and Karen
academic science." Minerva, 21: 198-233. Jones, L., G. Seashore Louis 1981 Stability and Change: Innovation in
Lindzey, and P. Coggeshall (eds.) 1982 An Assessment of an Educational Context. New York: Plenum. Seashore,
Research-Doctorate Programs in the U.S. Washington, Stanley E. 1954 Group Cohesiveness and the Industrial
DC: National Academy Press. Kanter, Rosabeth M. 1983 Work Group. Ann Arbor, MI: Institute for Social Research.
The Change Masters, New York: Simon and Schuster. Snow, Charles C., and Lawrence G. Hrebiniak 1980
Kerr, Jeffrey, and John M. Slocum 1987 "Managing "Strategy, distinctive competence and organizational
corporate culture through rewards systems." Executive, 1: performance." Administrative Science Quarterly, 25:
99-107. Krimsky, Sheldon 1984 Letter in geneWATCH, 1: 315-334. Van Maanen, John 1976 "Breaking in:
3. Liebert, Roland J. 1977 "Research -grant getting and Socialization to work." In R. Dubin (ed.), Handbook of
productivity among scholars: Recent national patterns of Work, Organization and Society: 67-130. Chicago: Rand
competition and favor." Journal of Higher Education, 48: McNally. Varrin, R., and S. Kukich 1985 Letter in Science,
164-192. Louis, Karen S., Melissa Anderson, and Judith P. 277: 385. Wade, Nicholas 1984 The Science Business.
Swazey 1988 "The university’s role in regulating graduate New York: Priority Press. Weick, Karl E. 1976 "Educational
education and research." Paper presented at the annual organizations as loosely coupled systems." Administrative
meeting of the Association for the Study of Higher Science Quarterly, 21: 1-19. Zuckerman, Harriet, and
Education, St. Louis, MO. Marsh, Herbert, and Katherine Robert K. Merton 1972 "Age, aging and age structure in
Dillon 1980 "Academic productivity and faculty science." In M.W. Riley, M. Johnson and A. Foner (eds.),
supplemental income." Journal of Higher Education, 51: Aging and Society: A Sociology of Age Stratification, 3:
546-555. Mazzoni, Tim L. 1987 "The politics of educational 292-356. New York: Russell Sage Foundation. [Tabular
choice in Minneasota." In William Boyd and Charles Data 1 to 5 Omitted] [Figure 1 Omitted]
Kerchner (eds.), The Politics of Excellence and Choice in
Education, 1: 109-141. London: Taylor and Francis.
McClelland, David C., John W. Atkinson, Russell A. Clark,
and Edgar L. Lowell 1976 The Achievement Motive. New
York: Irvington. Merton, Robert G. 1968 "The Matthew
Effect in science." Science, 159 (January): 56-63.
Mintzberg, Henry 1973 "Strategy making in three modes."
California Management Review, 16: 44-53. Pelz, Donald
F., and Frank Andrews 1976 Scientists in Organizations.
Ann Arbor, MI: Institute for Social Research. Peters, Lois,
and Henry Fusfeld 1982 "Current U.S. university/industry
research connections." In National Science Board,
University-Industry Research Relationships: Selected
Studies. Washington, DC: U.S. Government Printing
Office. Peters, Thomas J., and Robert H. Waterman 1982
In Search of Excellence. New York: Warner. Peterson,
Richard A. 1981 "Enterpreneurship and organization." In
Paul C. Nystrom and William H. Starbuck (eds.),
Handbook of Organizational Design, 1: 65-83. New York:
Oxford University Press. Peterson’s Guides 1984
Peterson’s Guides to Graduate Programs in the Biological,
Agricultural and Health Sciences. Princeton: Peterson’s
Guides. Pfeffer, Jeffrey, and William L. Moore 1980
"Power in university budgeting: A replication and
extension." Administrative Science Quarterly, 25: 398-418.
Pfeffer, Jeffrey, and Gerald R. Salancik 1974
"Organizational decision making as a political process: The
case of a university budget." Administrative Science
Quarterly, 19: 135-151. Ping Charles 1980 Industry and
the Universities: Developing Cooperative Research
Relationships in the National Interest. Washington, DC:
National Commission on Research. Ravetz, James 1971
- Reprinted with permission. Additional copying is prohibited. - GALE GROUP
Information Integrity

Vous aimerez peut-être aussi