Academy Management &Education, 2010, Vol. 9, No. 4, 638—651.

Does Business School Research

Add Economic Value for
Students?
JONATHAN P. O'BRIEN
Rensselaer Polytechnic Institute

PAUL L. DRNEVICH
The University of Alabama

T. RUSSELL CROOK
The University of Tennessee

CRAIG E. ARMSTRONG
The University of Alabama

The scholarly research conducted by business school faculty has long been the subject
of intense criticism for lacking relevance and value to practice. In contrast, we
theorize tha such research is relevant and valuable in that it contributes to what is
arguably the most critical metric of relevance for business school students: the
economic value they accrue from their education. We investigate this counterargument
on a sample of schools over an 8-year period. We find that research adds
significant value in that it can potentially enhance st ent salaries up to $24,000 per
year. However, we also observe that can lea to zmznzs Ing or even neaative
returns for students, and a research focus solely on elite journals might rob students of
the benefits of exposure to a broader array of new ideas.
Dierdorff, 2009), and according to Ghoshal (2005), may
The scholarly research activity conducted at ness schools busiof actually negatively influence practice.
has come under harsh criticism as late with some business Management Learning & Education for their comments, sugges- recruiters (Bradshaw, 2007). A recent survey by tions
practitioners even arguing that such research is a "vast Admissions
of Council
wasteland" irrelevance (Bennis & O'Toole, 2005: 99). et this
criticism IS not new. For almost 2 decades, schol 638
themselves have been expressing concerns that an
Copyright of the Academy of Management, (111 rights reserved. Contents may not be copied. emailed, posted to a li.stserv, or otherwise transmitte
excessive preoccupation with theory might bind business articles for individual use only.
school research into a "strai htjacket" that limits its Given these criticisms, in conjunction with the
relevance and value to practice e.g., Bettis, 1991; Daft &
centrality of research to most major business schools
Lewin, 1.990). However, these criticisms now seem to be
(Rindova, Williamson, Petkova, & Sever, 2005), one
reaching an almost feverish pitch, with many prominent
might assume that graduate business education is
scholars suggesting that research has overemphasized rigor
deeply troubled. Yet some evidence suggests
otherwise. Graduate business education has
experienced phenomenal growth over the last 3
The authors wish to thank Ben Arbaugh, Neal Ashkanasy, Don decades, with over 100,000 MBA degrees awarded in
Bruce, Duane Ireland, Michelle Kacmar, Dave Ketchen, Barry t
he United States in 2000 alone (Friga, Bettis, &
Mason, Bob Rubin, Donald Siegel, and severcd anonymous reviewers at
Sullivan, 2003; Morgeson & Nahrgang, 2008), and it
the Academy of Management and the Academy of has remained popular with both students and
(e.g., following the scientific method) and theory (Hambrj.ck, (GMAC) indicates that 94% of MBA
2007) at the expense of relevance and value to practice graduates lieve that their decision to
(Bartunek, 2007; Hambrick, 2007; McGrath, 2007; Pfeffer, pursue an MBA was the "right" 2006).
2007; Tsui, 2007). To make matters worse, such research Empirical evidence indicates that the
may be guiding what faculty teach in the classroom (Rubin &
2010 O'Brien, Drnevich, Crook, and Armstrong 639

research agendas of business scholars individual student's knowledge, skills,

are indeed shaped by the problems that and abilities, We also use a dynamic
the managers of large organizations panel data model that covers 658 gradu-ø
consider important (Barley, Meyer, & ate business schools from around the
Gash, 1988). Further, in their globe over an 8-year period to
assessment of whether MBA curricula is investigate the relationship between
aligned with managerial competencies, research and longer term economic value
Rubin and Dier*. dorff (2009: 22) found for students. Using this approach also
that ."the relevant training. allows us t more effectively control for
a myriad of potentially confounding
grounds [for managerial factors that could induce a spurious
competencies] are likel to be found relationship between research activity
in institutions of elevated and economic value accrual for students.
While no observational study can
researcha
definitively establish causality, our
approach and results doo
activity rather than institutions.where
research is de-emphasized and
provide strong evidence. that the
unsupported." Recent research has also
research con-
provided empirical support for the posi-

tive relationship between academic ducted by business school faculty

research an business school reputations does appear to
(Rindova et al., 2005), and rankings.
(Drnevich, Armstrong, Crook, & Crook, In benefit the students of their schools by
Press). Building on this evidence, we increasingø
contend that there is a vast gap between
popular perceptions and the reality of longer term salaries. Our evidence is
the relevance an value of business in direct contrast to the claims of
school research.(see Peng & Dess, 2010, others who lament the lack of relevance
for a review of the discourse on this and value of academic research.
issue). However, we also observe a note of
We theorize that the scholarly caution from our findings in that the
research activity conducted at business economic value created for students
schools in general will appears to diminish, and can even
contribute to what is arguably the turn negative, when a business
single most impo!tant metric of school's level of research activity
relevance for students: the economie becomes excessive, Somewhat similarly,
we also observe that students
salaries) they accrue from their benefit from
education, While recent studies have
found some empirical evidence of a exposure to diversity in research
relationship between research cmd MBA activities, in that some research
salaries (Friga et al., 2003; Mitra & published outside of the "A list
Colder, 2008; Morgeson & Nahrgang, 2008; jourrials" can also add significant
Rindova et al., 2005), there remains a value for students.
significant need for additional research
in this area. For example, while Rindova
et al. (2005) found that research has a
positive effect on the starting salaries
of MBA students, it is only an indirect THEORY AND HYPOTHESIS DEVELOPMENT
one, mediated by the reputation of the
Before we develop our arguments for the
school. Mitra and Colder (2008) also
relationship between the research
found that scholarly research has
conducted by business school faculty and
similar positive long-term effects on
student economic value creation, we
applicant, recruiter, and academic
first acknowledge that even if we find
perceptions as well as on performance.
no relationship between research and
However. the results of these two
value, this finding would not
studies are not surprising, given that
necessarily invalidate business school
both utilized measures based on MBA
research. The Merriam-Webster (2008)
starting salaries. Such measures are
dictionary defines a university as "an
problematic since new graduates are
institution of higher learning providing
unknown commodities, and it is
facilities for teaching and research
reasonable to expect that employers
Thus, even if we assume tha enhancing
would base new hire decisions largely on
economic value for students is an im-
the reputation of the graduate student's
institution (Rindova et al., 2005). In
contrast, we utiliz measures based on portant component of the teaching
average MBA salary appreciation 3 years objectives of business schools% the
after graduation,.which more acct- research function may still
t
yrately reflects he value of the
640 Academy of Management Learning & Education December

exist as an independent institutional that faculty consume when resources are

objective that does not need to bountiful. Such an interpretation would
contribute to an institution's teaching be consistent with Jensen's (1986)
objectives. That is, contention that agency costs are most
true to the basicresearch mission of acute when financial slack is abundant,
many universities (Scott, 2006), the or Bromiley's argument that excessive
generation of new knowledge may be a organizationa slack can lead t.
vital organizational objective, inefficient resource allocation
regardless of its immediate measurable decisions.
economic impact.1 Alternatively, faculty research
activity may simü ply represent
efficient resource allocation, Hosios
and Siow (2004) argued that faculty are
Of course, even if research is not motivated the closest thing that a university has
by financial returns, the knowledge emanating to residual claimants because increases
from it can lay the foundations for in faculty salaries have been driven
advancements of enormous commercial largely by the residual that is left
significance (Blaug, Chien, & Shuster, 2004), over after accounting for all fixed
sometimes years or even decades after the
research is conducted (National Research
costs and support staff. Since faculty
Council, 2004). generally take an active role in the
governance of the university, if their
Confounding Considerations research represented squandered
resources, then we would expect to see
Although pursuing scholarly academic
many instances of the residual claimants
research may be a legitimate objective
pressing for a more efficient resource
for business schools even in the absence
allocation model. For example, if.
of explicit financial returns, existing
faculty viewed research as wasteful of
empirical evidence does suggest that
university resources, we would then
research and economic value creation for
expect them uto seek to curtail new
students are at least correlated (e.g.,
hiring and increase teaching Odds of
Friga et al., 2003; Mitra & Golder,
existing faculty in return for
2008; Morgeson & Nahrgang, 2008; Rindova
commensuratelyv greater. pay, However,
et al., 2005). Casual observation would
since we observe little evidence of
also suggest that there is generally a
faculty pushing such initiatives, it is
positive correlation between research
plausible that faculty research activity
and business school rankings, even if
may be a product of the efficient
the rankings do not explicitly consider
distribution of the university's
or effectively measure research activity
resources. Although some may view the
(Drnevich et al.) In Press). Of course,
resources devoted to research as a
such correlations could well be spurious
lavish perquisite, it may sometimes be
(i.e., noncausal) if some unobserved
efficient for an organization to provide
other factor(s) was, in reality, the
lavish nonpecuniary benefits to
true causal driver of both research
employees (or managers) if the employees
activity and student economic value
will accept lower wages in return for
creation (or MBA rankings). While there
those benefits (Demsetz, 1983; Rajan &
are many possible factors that could
Wulf, 2004). Thus, if most faculty
confound this relationship, we speculate
members enjoy doing. research (or at
that a school's reputatione alumni
least enjoy it more than the other
network, and financial resources may be
components of their jobs, such as
the most likely candi;, dates, Of
service or teachIng), then by offering
course, these potential confounding
faculty more time and resources to
factors are not mutually exclusive.
conduct research, the university might
First, a school's reputational capital be able to recruit better faculty more
may help to attract both faculty and cost effectivel%. Hence, even if faculty
students of higher quak ity (Rindova et research did pot die rectly benefit
al., 2005). Second, a strong alumni students, _ studentsuwho« atten4
network may help the schoo p ace Dits research-intensive universities might
current students in better positions,oor indirectly reap more economic value than
with better employers, while also students who attend teaching-focused
allowing it to raise the funding ones.
necessary to provide faculty both the
time and resources to pursue research.
Third, independent of the alumni.etwork, The Positive Relationship Between
a school's financial resources students Business
an enriched learninc environment while School Research and Economic Value
also allowing nit to indulge faculty in
their desire to conduct research%ln the All of the above scenarios
latter two scenarios, the organizational describe why there might be a
resources devoted to faculty research spurious relationship between
might be considered a type of perquisite faculty research activity and
2010 O'Brien, Drnevich, Crook, and Armstrong 641

student economic value creation. rigorous approach to problem solving

However, if we can demonstrate a and decision making in the
relationship between research classroom. 1 Such rigorous solving
activity and economic value after might resonate with students, and
controlling for the most relevant they rmght make better decisions
confounding factors, then it would once they complet their programs.
provide strong evidence that the Similarly, business school researc
researc conducted by faculty at a generally emphasizes the contingent
given Institution does nature of relationships (e.g.,
generallybenefit the students that Lawrence & Lorsch, 1967). As a
attend that institution. It is our result, students of xesearch active
contention that such a faculty will likely be better
relationship does indee exist. prepared to tackle the complex
problems in business today. As
In order for research activity to
Morgeson and Nahrgang (2008: 12)
generate economic value for
aptly observe, "actively publishing
students, it must impart upon
_faculty can take their cutting-edge
those students valuable resources
knowledge to the classroom to
or capabilities not shared by all
enhance the learning of their
MBA graduates (Barney, 1991).
students, giving students a
Developed modern economies are
competitive advantage compared io
largely knowledge-based economies,
students m programs where there is
which indicates that knowledge-
not as much new nowledge.'
based advantages are important for
high performance (Miller & Collectively, these arguments
Shamsie, 1996). Correspondingly, suggest that students should stand
faculty who are_actively engaged to benefit from the scholarly
research activities of their
in research can likely provide
instructors. Thus, we believe that
value for their students y business schools with higher levels of
ransferring to them new knowledge research activity will create more
gleaned from their own research. economic value for their students than
In addition, even if an business schools with low levels of
individual faculty member's own research research activity. So long as labor
has little relevance to practice, being markets are at least semicompetitive,
actively engaged in research helps graduates of business schools with
faculty keep abreast of, and involved higher levels of research activity
with, "cutting edge" knowledge should be able to appropriate at least
developments in the field. Faculty in some of the economic rents attributable
turn can transfer suckl knowledge to the to their knowledge, skills, and
student through classroom interactions abilities (Coffl 1999). Stated
Therefore, students are most likely to formally:
gain knowle ge-based advantages from
faculty who active y engage with a Hypothesis 1: There is a
research community developing new positive
intellectual capital and who stay.
abreast of innovative developments in relationship
the fie d. In contrast, such knowledge- between the level
based gdvantages are unlikely to accrue
to students at business schools where of a business
faculty _are not actively engaged in school's research
suc}l research communities,
Further, although rigorous,
theory-driven research has been 1 This sentiment was also echoed in a recent
criticized by some for being too call for papers by The UNESCO Forum on
"scientific" (e.g., Bennis & Higher Education, Research, and Knowledge.
O'Toole, 2005), we think that One of the central themes of the colloquium
faculty and students exposed to the was that universities that lack the resources to
scientific process in such research sustain research programs may be limited "to
might also realize some benefits. In being-institutions pf knowledge dissemination
particular, active engagement in rather than knowledge creation".and that
knowledge creation through research, "[u]niversities that are deprived, or deprive
as opposed to simply teaching from themselves, of that ingredient risk the
intellectual erosion of their programs of study,
textbooks and educational materials lose their critical ability to assess claims to
that others write, may help faculty knowledgeeand become dependent on the
hone their analytical skills and, outside supply of knowledge ' (UNESCO,
consequently, emphasize a more 2006B
642 Academy of Management Learning & Education December

activity and the between the level

amount of of a business
economic value school's research
created for activity and the
students. amount of
economic value
The Curvilinear Relationship Between created for
Business School Research and
Economic Value students.
Although we believe that higher METHODS
levels of research activity will Sample
generally add economic value for
the students of a business school, We had two major considerations for
constructing the sample used to test our
we do not contend that research is
hypotheses. First, because the MBA is an
a continuous and inexhaustible
increasingly global product (Bradshaw,
lever for improving economic value 2007), we desired an internationally
for the students, In keeping with diverse sample of business schools. The
the law of diminishing returns second consideration was a more
(Malthus, 1815), we would expect practical one. Although there are
that at some point, additional thousands of schools offering MBA-type
units of research activity would degrees in the world, we needed to limit
provide negligible additional the sample to a manageable number of
benefits to students. Hence, the schools for which reasonably reliable
relationship between research data were available. Accordingly, we
chose to include in our sample all
activity and economic value
business schools that are members of the
creation for students should be Association to Advance Collegiate
relatively asymptotic. However, it Schools of Business (AÄCSB) and
seems likely that at some point, participated in its Annual Business
the relationship would actually School Questionnaire survey between the
turn negative. Almost all years 2001 and 2007. To construct the
organizations have limited full sample (described in detail below),
resources available to pursue their we combined the AACSB data with data
objectives (Barney, 1991), and the from the Financial Times Global MBA
individuals within those Rankings (henceforth FT Survey).for the
years 2002 to 2009, and with publication
organizations have limited
count data derived from the Institute
cognitive resources to devote to for Scientific InformaWeb ofknowledge
multiple tasks (Norman & Bobrow, database for the years 1999 through
1975). An excessive focus on one 2006. Our final sample encompassed 658
activity must, at some point, come different&usinessø schools and, 3482
at the expense of other activities ,total observations„
and objectives (Kaplan & Norton,
1992). Thus, if a school places an
excessive focus on research, here
might beless focus placed on Variables
teaching and knowledge transfer. As Dependent Variable
a resulte faculty will in; vest
As we are interested in investigating
relatively less effort in teaching
whether the research conducted by
and studen outcomes business school faculty adds economic
wilUsuffer..Thus, while the value for students, we measure this
research activities of business value by way of the average percentage
school faculty might generally increase in MBA salary (versus pre-MBA
benefit students, excessive salary) 3 years after graduation jor
preoccupation with research will thegverage alumnlvof a schoo Measuring
likely attenuate this relationship. percentage increase as opposed to simÉly
Thus: measuring the raw salary 3 years after
graduation helps to control forGncoming
Hypothesis 2: There is a studen! quality. and also allows us to
curvilinear assess how much the student, actually
benefited from the MBA experience
(inverse Ushaped) Further, ong-term salary improvement is
relationship also arguably one of the most important
considerations for business school
2010 O'Brien, Drnevich, Crook, and Armstrong 643

applicants. We collected the data for information , Specifically, we gathered

this variable directly from the data on the impact factor ratings for
Financial Times MBA rankings for the the years 2003, 2005, and 2007 for the
years 2002 to 2009. To the best of our journals in our subject areas of
knowledge, the FT Survey provides the interest. For each journal, we averaged
most comprehensive and thorough these three ratings, then sorted the
information publicly available on list and categorized the top-40 as "A
postgraduation MBA salaries. However, as journa s" (see Appendix). For every
this variable is only available for the article appearIng in one of these 40
schools ranked by the FT Survey in any journals, we assigned fractional score
given year, it is not available for based oyßhe number of authors (e.g., an
every observation in our sample (we article with oneAthorÄs scored as a l,
address this issue below in the where an article with two authors is
Analytical Considerations section). scoredgs 0.5, etc.), then ye computed a
Finally, it is worth noting that the (based on author affiliations) for ever year.
salary data reported by the FT Survey Our final measure for this level of research
(and used in our analysis) are moving activity,
averages, which take into account the A-publications, is this sum divided by
previous 1—2 years of surveys to smooth the total number of full-time equivalent
out the effects of shortterm faculty members.
aberrations. This smoothing should not Aligning the salary data with the
bias our results in favor of our publication data presented some special
hypotheses, as it will only reduce considerations. The FT Survey publishes
variability in our dependent variable, its data early in the year (generally
making it more difficult to find January) from surveys conducted the
statistical significance. previous year of graduates from 3 years
prior. Furthermore, as the FT Survey
uses moving averages to smooth out
Independent Variables aberrations, the salary data for a given
year may encompass students that
We measure the level of research actually graduated 6 years prior. For
activity of a business school with the example, the 2007 salary data consists
vanable publication. whicWwe of surveys done in 2006, mostly to
constructedfrom dataderived trom the graduates from 2003, but possibly
ISI's Social Science Citation Index, We encompassing graduates from 2002 and
began by downloading all citations for 2001. Further complicating matters is
the years 1999 to 2006 for all journals the fact that active and even relatively
listed under the subject areas of mature research projects going on in
"business," "business, finance," (for example) 2002 may not actually
'management," and "operations research & appear in journals (i.e., be published)
management science." After excluding for several years. To account for these
some nonacademic journals (e.g. Forbes, lags, we matched salary data to the
Fortune), our data encompassed 254 average number of publications 3 and 4
jour-. nals and 45,325 unique Journal years prior (e.g., we merge salary data
articles with 77,977. for 2007 with the average number of
•neo
[nonunique] authors (i.e., an average of publications in 2003 and 2004). As a
1.7 authors. pey paper). We then robustness check, we also adjusted this
constructed the variable publications in lag structure a year or two in each
three different ways. First, we direction and found that doing so does
constructed a measure that only included not materially alter our results.
pub ications In elite "'A list upon As a further robustness check,ye also
their reputa% tion for rigor and examined. the possibility that research
_impact. .According to Rindova and activity published in journals outside
colleagues (2005), the norms of modern of the "A jouynals" might also. create
science prescribe that high-visibility value for stu ents. To examine this
publications such as these serve as the possibility, we constructed the variable
most effective institutional B-publicationso in the same manner as A-
certification of a faculty's research publications. Specifically, when we
quality. Thus, we used ISI's Journal ranked journals based on their average
Impact Factor ratings to develop a list ISI Impact Factor, we classified all
of the top-40 journals in business. journals that ranked between number 41
Although impact factor ratings are an and number 120 as B-)ournals Gee
imperfect measure of journal quality, Appendix). Finally, we also constructed
our approach is objective and our a third measure of research activity, C-
derived list of journals corresponds publications, which encompassed all
closely to more subjective journal lists remaining journals. In order to test the
used by the FT survey and the University curvilinear (inverse U-shaped)
of Texas at Dallas' Top-IOO Business relationship we predict in Hypothesis 2,
School Research Rankings (see http:// we also include the square of each of
top100.utdallas.edu/ for more the publication variables in each model.
644 Academy of Management Learning & Education December

Control Variables of the sample). Further, we also

included a dummy variable indicating 9
In order to isolate the effect of the
school had a doctoral program. We
research conducted by business school
derived data lor this item from the FT
faculty on economic value creation for
Survey and from the AACSB data. In
students, we explicitly controlled for
instances where this data item was
factors that could In uce a spurious
missing, we supplemented the data
relationship between faculty research
either by performing web searches,
and studen outcomes We use the variable
contacting the school, or by authors'
budget to measure the school's
judgments based on familiarity with the
operating budget per fulEtime faculty
schools.
member to control for the financial
resources of the school. We believe Unfortunately, some constructs which
that the operating budget is a more we considered including, such as
pertinent measure of a university's "quality of the alumni network," are
financial resources than alternative difficult, if not impossible, to
measures such as endowment because many measure. Fortunately, using a dynamic
schools (especially non- panel data model allows us to introduce
U.S. schools) are quite competitive an additional level of implicit control
despite having little to no endowment. for potentially confounding factors
Further, in terms of both research such as these (Wooldridge, 2003). To
intensity and student outcomes, how the extent that unmeasured factors such
much the school is actually spending on as "the quality of the alumni network"
its operations should be more important are relatively stable over time,
than the size of a school's invested including the lag of the dependent
endowment. We also controlled for the variable in the model helps to control
size of the faculty by incliiding the for the extent to which such factors
number bf full-time equivalent (FTE) might affect the values of the
faculty at each school. We further dependent variable, salary Ssee
controlled for the tuition charged by Wooldridge, 2005).
the sGhool, since fhe cost of the
education affects botß the school's
finances and the overall value for the Analytical Considerations
money for the student. We constructed
the variables budget, faculty, and Panel Data
tuition from datg derived from the
AACSB's, DataDirect service Testing our hypotheses presented some
(AACSB/2009). critical analytical considerations.
We also control for a school's First, we have
reputation (Rindova et al., 2005; multiple ohservations school.
Saf6n, 2007) by constructing a measure (i.e., panel data). With panel data, if
based on recent rankings. As the FT we fail to control for unobserved factors
Survey did not rank many schools in our that are correlated with both our
sample, but points out that tiers are dependent variable and one or more of our
probably much more meaningful than a independent variables, then traditional
school's absolute rank, we constructed ordinary least squares (OLS
dummy variables corresponding to regression will produce biased results
ranking tiers. hust we measure the because the afflicted independent
variable Tier 1 as one if the FT Survey variable(s) will be
ranked the school in the top-50 within correthewfrror
the last 2 years, and zero otherwise. (Wooldridge, 2004). Therefore,
Similarly, we measure Tier 2 as one if all models employ random effectsofor the
the FT Survey ranked the school between schools.
51 and 100 within the last 2 years, and
zero otherwise. A third tier, which
serves as the default condition,
Censored Data
encompasses all schools not ranked in
the last 2 years by the FT Survey. As a In addition to having multiple
robustness check, we also experimented observations per school, some data
with more tiers. Using more tiers did items were simply missing, as the FT
not change our results, nor did it Survey did not rank every AACSB
materially improve model fit. Thus, we school in our sample in a given
chose to present the more parsimonious
year. This issue presents another
three-tier classification system.
critical methodological concern
We also included in the analysis a because the data are not just
number of additional Other-NA, Europe,
missing at random, but are
and Asia) to control for the geographic
location of the school. The default—
location. would hence be anywhere censored due to the school falling below
outside these areas (approximately 5% a certain threshold, In the •FT
Survey, the percentage increase in
2010 O'Brien, Drnevich, Crook, and Armstrong 645

salary, along with the highly correlated

gross weighted total salary, are by far Hypothetical Example: Censored
the most heavily weighted components of Dependent Variables
the rankings. Thus, when the salary data
are missing, it is safe to assume that
in general the school fared toward information from the observations with
missing values of the dependent variable
the low end of the distribution on this item.if observations that fall above and
Therefore, estimating a conventional e.g.,below the censoring threshold have
OLS) model on censored data will result insystematic differences in the
inaccurate estimates because the sample isdistributions of their independent
biased and not necessarily representative ofvariables. The two most common methods
the full population (Wooldridge, 2003). As afor making use of the information
stylized example, consider the data depictedinherent in the observations that fall
in Panel A of Figure l. below some censoring threshold, and
If we could measure all individuals in thereby correcting the bias introduced
the hypothetical population in Panel A, by the censoring, are the Heckman model
we would find a correlation between the and the Tobit modeleThe Heckman model
two variables of approximately 0.48. iso most appropriate when the censoring
However, if we restricted our sample to is due to the. endogenous self-choice of
just high-performing observations (i.e.,
those scor-
ing above 70), we observe a correlation
of only 0.026. As we illustrate in Panel the subject>, such as when individuals
B of Figure l, a censored sample might decide not to work because the available
be even more problematic if the true wage falls below their yeservation wage
underlying relationship is nonlinear. (Heckman, 1979). Conversely, is more
appropriate when the dependent variable
Despite this limitation, we can obtain
could potentially fall below the
valuable
censoring threshp do(unlike hours
A 100 worked, which cannot be negative),
but we simplvannot observe it (Maddala,
1991). Since we expect that all schools
would likely prefer to rank in the FT
Survey, and do in fact have some
[unobserved] score for salary, we deemed
the Tobit model to be more appropriate
for our data. We Winsorized the variable
salary at the 1st percentile of its
distribution and treated this value as
the censoring threshold.

Endogeneity
o 0.25 0.5
0.75 1 Unobserved heterogeneity presents a
problem with our data because not only
Independent Variable do we have multiple observations per
school, but also because some third
B 100 factor(s) could influence our prime
theoretical variable, publications, as
well as influence salary. Fortunately,
using panel data„presents ethe
opportunity to help omitted variables
bias. Hence, we
address the potential endogeneity issue
by using the dynamic Tobit model
Wooldridge (2005). This model allows us
to implicitly control for the numerous
unobserved quality) that might
influence salary by including a lag of
the dependent variable as a predictor
-variables Although introducing the lag
o 0.25 0.5 of the dependent variable as a
0.75 1 predictor variable can create bias,
Wooldridge (2005) explains that a
Independent Variable dynamic Tobit model that effectively
FIGURE 1 controls for unobserved heterogeneity
can be estimated by simply including
646 Academy of Management Learning & Education December

the initial condition of the dependent effective control for unobserved

variable in a random effects model confounding variables.
(i.e., the value in the first time In Model 2, we present the dynamic
period), as well as some time invariant Tobit model that incorporates all the
controls, including a vector of all the variables suggested by Wooldridge (2005)
values over time of a time-varying to control for bias and unobserved
variable. These, variables will absorb heterogeneity. For
the correlation with the unobserved highly
effects,0 permitting accurate estimates significant Chi-square statistics (p <
of relationships among the remaining 0.01), Indicating satisfactory model
variables. Hence, we include fit. As estimates for

as additional control variables the time constant variables may be

value of salary unreliable (Wooldridge, 2002: 541), we
refrain from interpreting those
in 2002 and a vector of the values for coefficients. However, the strong
Tier 2 for each school for the years significant coefficient on salary (2002)
2003—2009 implies that there is substantial
RESULTS he,Y1itial condiö tionßnd unobserved
heterogeneity. (Wooldridge, 2005).
We provide the descriptive statistics of In regard to our hypotheses, the
our sample in Table I and the results of significant positive coefficiento(p <
our statistical analyses in Table 2. 0.01) for A-publications in Model 2
Model I presents a standard (static) indicates support for -Hypothesis that
cross-sectional ran om effects Tobit e
ihere is a positive relationship between
model. However, we only present the leve of researchuactivityu and
purposes and refrain _$rom drawing student economigcvalue. Likewise, the
inferences beo significant negative coefficient (p <
0.01) for A-publications Model 2
cause it is static and does pot include indicates support for Hypothesis 2, that
the lag of the the relationship between the level of
research activity an student economic
dependent variable, yhich can help value is curvilinear (inverse-U). Taking
serve as an the first derivative o salary with
respect to publications
TABLE 1
Descriptive Statistics

Note: For Salary, statistics are based on 578 observations. For all other variables, n 3753.
Salary-2002 0.24 0.24
Tier 2 - 2003 2.00 1.66 1.00 2.07

Tier 2 - 2004 6.54 5.87 5.78 6.60

Tier 2 - 2005
-3.95 -3.93 -3.19
Tier 2 - 2006 4.13 3.34 3.90 3.66
Tier 2 - 2007 5.34 5.83 5.99 5.11
Tier 2 - 2008 0.75 0.53 0.80 1.08
Tier 2 2009 2.02 3.11 2.40 2.36
Observations 3753 3482 3482 3482 3482
Wald Chi-square
2010 O'Brien, Drnevich, Crook, and Armstrong 647

Inflection Point 0.17 0.16 0.09 0.10 .15(A); .08(B) Max. Value 40.7 15.1 15.2 11.5 IO(A); Il(B)

TABLE 2
All models also included year fixed effects (not reported for brevity).

< 0.10; < 0.05; < 0.01.

Statistical Results of the Tobit Models

Model 4 Model
Model 1 Model 2 Model 3
5
Constant and solving for the inflection point reveals that the journal" publications
24.1 24.1 ** 20.9*
do indeed
-10.0
add significant maximum value on the curve occurs at 0.16 publi- value over ana above 137.5*
A-Publications
journal" publi-
A-Publications2 461.0*
B-Publications
B-Publications2 generally'ail%to do sq, Moreover, because of the
C-Publications curvilinear relationships, a mix of-1.3
A and B pub-
C-Publications2 -1189.8±
Budget 3.E-04* 0.00 0.00 0.00
Tuition 0.03 0.00 0.00 0.00
Faculty
0.08
Tier I 26.77 26.51
Tier 2
49.40" 19.48
U.S. 15.66 -4.21 -7.88 -2.87 -8.73
Other-NA 13.71 -15.20 -11.70 -16.00±
Europe 32.27* 7.52 1.19 5.76 2.78
Asia -36.53* 13.20 10.96
Doctoral 6.91 6.78* 6.88* 6.29*
Lag-Salary 0.23
of the variable A-publications (the maximum value

catlons per FTE faculty member per year, which cations, but third-tier g..journal—
epubiicationsz

"B" and CO journals also creates value or stu- Conversely, 0.23 A-publications and
zero Bdents. Models and 4 produce relatively similar publications produce ct
predicted increase of just results to Model 2
using publications data based 7.3%. This suggests that a moderately broad yie svy

solely on second- and third-tier journals, respec- of academic scholarship, which can
theoreticallyo tively. In order to ascertain whether each tier of encompass g,yider
array of new and sometimes publication adds value for students while control- even
moFe.ygdical go

ling for other tiers of publications, we

include all three tiers in Model 5 (we
do not include the square of C-
publications because it was
insignificant and including it increased
potential collinearity problems). This
model suggestsu!hat second-tier "B
students.
648 Academy of Management Learning & Education December

In order to facilitate the publications would,equate to a gain of

interpretation of our results, we plot over $24,000 per year. This result
in Figure 2 the total contribution to strongly suggests that research-
salary at varying levels of A- and B- intensive schools generally do

0.05 0.1 0.15 0.2

Publications
FIGURE 2
Impact of Faculty Research on Student Salaries. The x-axis plots Publications (per FTE faculty per
year) from the 1st to the 99th percentile. (Note: the distribution of Publications varied slightly
across journal tiers), while the y-axis depicts the total increase in student salary (%) associated
with the corresponding level of publications. We derived predicted values from Model 5 of Table 2.
level publications, as suggested by through their future employment.
Model 5. We plot publications (per FTE From these results, one might
faculty per year) from the 1st to the conclude that the actual state of
99th percentile. However, publications
the relevance of business schoo
per FTE faculty per year can be
difficult to interpret, as many schools research is not nearly as dire as
employ numerous adjunct and clinical or the common perceptions some haves
professionally qualified faculty, and suggested. (e.g., Pfeffer & Fong,
may also have numerous academically who 2004; Skapinker, 2008). However, we
are no longer activel engaged in should also keep in mind that an
research. Further, the distributions of excessive focus on researc activity
publications are highly skewed. Hence, might starVto erode that value
in order to facilitate the premium%and that the students of
interpretation of our results, we use
dotted lines to designate the 95th to those schools.ppearing on
the 99th percentiles of publications diminishing returns side of the
ntriguingly, this plot suggests that inverted-U curv,e would likely
Bpublications can add comparable value benefiturom loweraevels of research.
to A- excessive levels of
publications, although the peak of their research activity attenuate student
value contribution occurs at a lower returns, more research-
level than Apublications, and they intensive schools still generally
experience more sharply declining produce better returns for students than
returns to additional publications. less research-intensive schools. For
example, even at the 99th percentile of
both A-publications and B-publications,
DISCUSSION student salaries reap a predicted
increase. Conversely, at the 25th
Our results indicate that the Level o
percentile of these distributions (which
scholarly re-. search activity at
equates to zero publications of each
business schools appears to add
type), no value is added for student
considerable economic value to MBA
salaries, ceteris paribus.
students' f ture salaries. Given that
the average MBA salary for programs The focus of our study was the general
ranked by the 2009 FT Survey is a!most relationship between the level of
$115,000 pe5eyear, the 21% predicted research activity in a business school
increase in salary that is possible at and student economic outcomes.
the optimal levels of both types of Accordingly, in our analysis we
2010 O'Brien, Drnevich, Crook, and Armstrong 649

examined the relationship between time. Hence, the results of our study
research publications and student have implications for comparisons both
salaries, and considered both variation across schools and longitu-
across schools and within schools over
a superior job in helping their students may not be generalizable to the entire
acquire and hone the knowledge, skills, population of business schools.
and abilities, which pays financial returns Futuregesearch.uld address some of
to the students dinally within schools, To these potential limitations by
the extent that some extending our studyén nymerous ways.
$irst,a it might prove enlightening to
schools exhibit significant variation examine whether the research activities
over time in their level of research of business school disciplines add more
activity, our conjecture is that general value than the researchugc$iyitiessof
trends are much more consequential for othev.discipIines.oAs the field of
student economic outcomes than year-to- management tends to be much more
year fluctuations- i n , publication. interested in theory than some other
countsvlndeed, although post hoc disciplines (Hambrick, 2007), such an
analyses suggested there is more analysis might yield useful fodder for
variation in,research activity across the current debates regarding the
schools than there is within schools theoretical straightjackets within
over did, exhibit a significant negative management (Bartunek, 2007; Hambrick,
trend in publicatiqp counts (using the 2007; McGrath, 2007; Pfeffer, 2007;
total of A- and B-publications), while Tsui, 2007). Second%it might also prove
over 100 schools exhibited a significant useful to differentiate the value added
positive trend (p < 0.05). In light of by normatiye versus positivist yesearch
the illustrative results of this simple activities. Speculatively, if
post hoc analysis, it appears an upward academically qualified business school
trend in research activity may likely be faculty are to add value in excess of
beneficial for the students at most of what students may acquire from
the schools in our sample, while a professionally qualified academic staff
downward trend might actually be more (e.g., clinical and adjunct faculty),
beneficial for students at schools in then it may be important to focus and-
only the upper echelons of research medical- faculty), more how things
activity. should be done, as opposed to how
things are actually done in practice.

Limitations. Implications, and Future

CONCLUSIONS
Research
Our objective here was to investigate
While our results have substantial
the apparent gap between the popular
practical implications for business
negative perceptions of the relevance
schools and their stakeholders, we do
and value of the rigorous theorydriven
stress some important caveats. irst,.
research activity conducted at business
while we demonstrate that scholarly
schools and the reality of the anecdotal
research activity generally adds
(e.g., continued popularity and growth
economic value for students, we cannot
of business school education) and
directly address whether this value is
empirical evidence to the contrary
attributable Dpurel to rigorous, theory-
(e.g., Friga et al., 2003; Mitra &
drivelA scholarly research, We can,
Colder, 2008; Morgeson & Nahrgang, 2008;
however, likely assume that our measure
Rindova et al., 2005; Peng & Dess,
of A-publications, based on the top-40
2010). To this end, we first theorized
leading business journals, is reflective
that such research activity, in general,
of the same research priorities of the
should likely contribute to what is
business schools that have drawn so much
arguably the single most important
criticism (e.g., Pfeffer & Fong, 2004;
metric of "relevance" for students: the
Skapinker, 2008). Second, we cannot
economic value they accrue from their
directl address whether business schools
education. We then utilized a dynamic
could have added even more value for
panel data model on a sample of 658
their students if they emphasized
business schools over an 8-year period
practical relevance fin gesearch and
to empirically examine this
teaching) over rigorous, theory-drivery
relationship. Our study offers both a
scholarlvesearchi we can onlwobserve
response to some of the critiques of
thatstudents appear to benefit
business schools (e.g., Bartunek, 2007;
economically fron»business schools with
Bettis, 1991; Hambrick, 2007; Pfeffer &
more active scholarly research_programs.
Fong, 2004; Pfeffer, 2007; McGrath,
Third, while we conducted our analysis
2007; Skapinker, 2008; Tsui, 2007), and
on a large sample that is representative
a validation and significant extension
of AACSB-accredited busines? schools,
of some prior related research (e.g.,
there are thousands of schools offering
Drnevich et al., In Press; Friga et al.,
MBA-type degrees, and.hus„our results
2003; Mitra & Golder, 2008; Morgeson &
650 Academy of Management Learning & Education December

Nahrgctng, 2008; 2008; Rindova et al.,

2005; Peng & Dess, 2010). First, in
response to the critiques and
commentaries, we find evidence that the
research conducted at business schools
is relevant and valuable to
practitioners as evidenced by the
considerable longer term economic value
added to MBA student salaries.
.ln,.terms of validation_apda extension
of prior research, we measured.co-,z
nomic value using salaries 3 years after
graduation gas opposed to using MBA
starting salaries), which is more
reflective of the value of the
knowledge, skills, and abilities
provided by the>usiness schoo education.
Additionally, we employed a dynamic
panel data model, allowing us to
implicitly control for a myriad of
potentially confounding factors for
which we could not measure explicitly
and control that could induce spurious
relationships in studies using
traditional models or cross-sectional
data. While only a controlled experiment
with random assignment can definitively
establish causality, our work does
provide strong evidence of the
probability of this causal link.
While our findings indicate that the
scholarly research activities of
business school faculty do indeed appear
to benefit students economically by
significantly enhancing their longer
term salary appreciation, we also
observe a note of caution. The economic
value created for students appears to
plateau, or can even diminish, when the
level of research activity becomes too
excessive or is overly restrictive.
Thus, while more research activity in
both "A" and "B" journals would appear
to be
2010 O'Brien, Drnevich, Crook, and Armstrong 651

better for most schools, there appears

to be diminishing returns from research
that can even turn negative from
excessive levels of research activity.
These observations would appear to be
useful for. deans and business school
administrators in eval; uating their
schools' positions on research weigh!
ing and in making decisions about
resource allpcations and incentive
systems.$inally, the results of this
study offer evidence to support the
recent conclusions of Ferris, Ketchen,
and Buckley (2008: 743) that perhaps
"we need to stop hand wringing and
apologizing for being organizational
scientists, and instead focus on

APPENDIX
Listing of 'A' and 'B' Journals Rank-Ordered by Average Impact Factor
Rank Tier Journal Avg. IF Rank Tier Avg. IF
Journal Avg. IF Rank Tier Journal
A MIS QUART
A ACAD MANAGE REV
A J RISK UNCERTAINTY
ACAD MANAGE J
A J RETAILING
J MARKETING RELIAB ENG SYST SAFE
MARKET SCI BRIT J MANAGE 1.327 81
A
ACCOUNT ORG SOC
1 J FINANC 4.538 INT J ELECTRON COMM 1.309 82 0.831
41 ENTREP REGION DEV
2
A J ACCOUNT ECON 4.347 WORLD BANK ECON REV 1.291 83 0.823
3
42 B 1.287 84 TRANSPORT RES E-LOG
ADMIN SCI QUART 3.520 J INF TECHNOL B
0.821
4 3.498
B 1.268 85 J WORLD BUS 0.820
A J FINANC ECON B J ORGAN BEHAV 1.263 86
B
5 3.217 45 B NEW TECH WORK EMPLOY 0.807
ORGAN SCI 3.056
B
MANAGE 1.228 87 0.800
6
STRATEGIC
46 B B COMPUT OPER RES
A 2.918 J PUBLIC POLICY MARK 1.185 88 B
7
MANAGE J
47 B
EXPERT SYST APPL B
J FINANC INTERMED 0.793
8 2.784 48 B 1.160 89 QUEUEING SYST 0.789
B INFORMS J COMPUT B
9 A J MARKETING RES 2.699 49 1.143 90
J CONSUM AFF 0.768
B B
10 J CONSUM RES 2.497 50 DECISION SCI 1.127 91 B 0.767
A J ADVERTISING
11 2.483 51 B B 0.758
REV FINANC STUD B DECIS SUPPORT SYST
A 1.127 92 B
RES POLICY
2.164 52 MATH FINANC PSYCHOL MARKET 0.757
12 1.124 93
J ACCOUNT RES 2.161 53
B
OPER RES B GROUP DECIS NEGOT 0.753
13 B 1.119 94 B
14 A J MANAGE 2.084 54 B ORGANIZATION 1.092 95 B
MANAGE LEARN 0.753
15 MANAGE SCI 1.817 55 B CALIF MANAGE REV 1.068 96 B IND MARKET MANAG 0.746
16 J OPER MANAG 1.758 56 B LONG RANGE PLANN 1.067 97 B TECHNOL FORECAST SOC 0.737
1.729 B B
17 HUM RESOUR 57
B
J FINANC QUANT ANAL 1.064 98
B
INT J OPER PROD MAN 0.733
18 A 1.689 58 FINANC MANAGE 0.724
MANAGE 1.688 59
B 1.060 99 B NATL TAX J
0.714
19 J BUS 1.018 100
ORGAN STUD 1.654 60 B B J BUS RES 0.678
20
A J INT BUS STUD B
J GLOBAL OPTIM
INT J MANA G REV 1.017 101 B
21 INFORM & MANAGE 1.651 61 B INT J RES MARK 1.014 102 B
J ADVERTISING RES
0.673
22 ACCOUNT REV 1.642 62 B B 0.667
LEADERSHIP QUART 1.641 63 B J QUAL TECHNOL 1.008 103 B NETWORKS 0.656
23 A
1.625 64 B INT J SELECT ASSESS 0.976 104 B TOURISM MANAGE 0.654
24 ORGAN RES
METHODS B MATH OPER RES 0.976 105 B
25 A 1.609 65 GROUP ORGAN 0.653
J BUS VENTURING 1.591 66 MIT SLOAN MANAGE 0.960 106 0.651
26 B B MANAGE
ORGAN BEHAV B REV
0.957 107
B
REAL ESTATE ECON
27 A 1.524 67 0.644
28 HUM 1.516 68 AM BUS LAW 0.933 108 B
WORLD ECON 0.643
B
DEC J HUM RELAT
29 1.506 69 0.925 109 B FINANC ANAL J 0.635
TRANSPORT RES B
J CORP FINANC INT J SERV IND MANAG
B-METH
30 1.499 70 B 0.921 110 B 0.628
J MANAGE J MONEY CREDIT BANK B J INT MONEY FINANC
1.452 n B B 0.610
31 INFORM 72 B INT J FORECASTING 0.920 111 B
J OPTIMIZ THEORY APP
32 SYST 1.431 B
0.607
A 73 J IND ECON 0.913 112
33 1.421 J BANK FINANC 0.606
J MANAGE STUD 1.385 74 DISCRETE EVENT DYN S 0.884 113 B
BUS HIST 0.605
34 B
0.877 114
J MONETARY ECON TRANSPORT SCI J INT MARKETING
MATH PROGRAM 75 B
35 1.375 B 11E TRANS 0.604
ACAD MANAGE J MANAGE INQUIRY 0.876 115
76 B
36 EXEC 1.375 B INTERFACES 0.601
J SCHEDULING 0.851 116
B
B
37 J PROD INNOVAT 1.374 0.595
MANAG 78 OMEGA-INT J MANAGE B J OPER RES SOC
38 1.374 B 0.844 117 B 0.593
J ENVIRON ECON S
79 B 0.843 118 AUDITING-J PRACT TH
39 MANAG 1.366
80 J ECON MANAGE MARKET LETT 0.587
40 CORP GOV 1.329 0.842 119 0.585
STRAT PROBAB ENG INFORM SC
SYST CONTROL 0.836 120
EUR OPER RES J BUS ETHICS
LETT
HARVARD BUS REV IEEE T ENG MANAGE
I ACAD MARKET
SCI
652 Academy of Management Learning & Education December

pushing knowledge and applications in

this field forward in meaningful ways."
2010 O'Brien, Drnevich, Crook, and Armstrong 653

REFERENCES Management Learning & Education, 4:

AACSB. 2009. The Association to Advance 75--91.
Collegiate Schools of Business. GMAC. 2006. MBA Alumni Perspectives
http://www.aacsb.edu/knowledgeservices/
Survey. Graduate Management Admission
datadirect/
Council - Comprehensive Data Report.
Barley, S. R., Meyer, G. W., & Gash, McLean, VA: GMAC.
D. C. 1988. Cultures of culture
Hambrick, D. C. 2007. The field of
Academics, practitioners ctnd the
management's devotion to theory: Too
pragmatics of normative control.
much of a good thing? Academy of
Administrative Science Quarterly,
Management Journal, 50(6): 1346-1352.
33(1): 24-60.
Heckman, J. 1979. Sample selection bias as a
Barney, J. B. 1991. Firm resources and specification error. Econometrica, 47: 153--161.
sustained competitive advantage. Journal
of Management, 17: Hosios, A. J. , & Siow, A. 2004. Unions
without rents: The curious economics of
Bartunek, J. M. 2007. Academic- faculty unions. Canadian Journal of Economics,
practitioner collaboration need 37: 28-52.
not require joint or relevant Jensen, M. C. 1986. Agency costs of free
research: Toward a relational cash flows, corporate finance, and
scholarship of integration. takeovers. American Economic Review, 76:
Academy of Management Journai, 32&-329.
50(6): 1323-„1333. Kaplan, R. S., & Norton, D. P. 1992.
Bennis, W. G., & O'Toole, J. 2005. How Balanced scorecard: TransIctting
business schools lost their way. Harvard strategy into action. Harvard Business
Business Review, 83(5): Review, Jan Feb: 71-80.

Bettis, R. 1991. Strategic management and the Lawrence, P. R., & Lorsch, J. W. 1967.
straightjacket: An editorial essay. Organization Differentiation and integration in complex
organizations. Administrative Science
Science, 2: 315819.
Quarterly, 12: 1—47.
Blaug, S., Chien, C., & Shuster, M. J. 2004. Maddalct, G. S. 1991. perspective on the use
Managing innovation: University-industry of limiteddependent and qualitative
partnerships and the licensing of the variables models in accounting research.
Harvard mouse. Nature Biotechnology, The Accounting Review, 66:
22(6): 761—764.
Malthus, T. R. 1815. The grounds of an
Bradshaw, D. 2007. A new dawn for opinion on the policy of restricting the
the MBA. Financial Times, January importation of foreign corn. The Works of
Thomas Robert Malthus, E. A. Wrigley cmd
29: I. D. Souden, (eds.), 1986, 7: 151—74.
Bromiley, P. 1991. Testing a causal model of London: William Pickering.
corporate risk taking and performance.
Academy of Management Journal, 34: 37-59. McGrath, R. G. 2007. No longer a stepchild:
How the management field can come into
Coff, R. W. 1999. When competitive advantage its own. Academy of Management Journal,
doesn't lead to performance: The resource- 50(6): 1365...1378.
based view and stakeholder bargaining
Miller, D., & Shamsie, J. 1996. The resource-based
power. Organization Science, 10(2): view of the firm in two environments: The
Hollywood film studios from 1936 to 1965.
Daft, R., & Lewin, A. 1990. Can organizational Academy of Management Journal, 39(3): 519—543.
studies begin to break out of the normal
science straightjacket? An editorial Mitrct, D.. & Golder, P. N. 2008. Does
essay. Organization Science, l: 1—9. academic research help or hurt MBA
programs? Journal of Marketing, 72(5): 50—
Demsetz, H. 1983. The structure of ownership 68.
and the theory of the firm. Journal of Law
Morgeson, F. P., & Nahrgang, J. D. 2008.
and Economics, 26: 375—390.
Same as it ever was: Recognizing
Drnevich, P., Armstrong, C. , Crook, T. A. , & stability in the Business Week rankings.
Crook, T. R. In Press. Do Research and Academy of Management Learning &
Education Matter to Business School Education, 7: 2æ.41.
Rankings? International Journal of
National Research Council. 2004. Science,
Management in Education.
medicine, and animals. Washington, DC:'
Ferris, G. R.r Ketchen, D. J. , Jr., & The National Academies Press.
Buckley, M. R. 2008. Making a life in the
Norman, D. A., & Bobrow, D. J. 1975. On data-
organizational sciences: No one ever said
it was going to be easy. Journal of limited and resourcelimited processes.
Organizational Behavior, 29: 741-753. Cognitive Psychology, 7: 44—64.
Friga, P., Bettis, R. , & Sullivan, R. 2003. Peng, M. W., & Dess, G. G. 2010. In the
Changes in graduate management education spirit of scholarship. Academy of
and new business school strategies for Management Learning & Education, 9: 282
21st century. Academy of Management 298.
Learning & Education, 2: 233—249.
Pfeffer, J., & Fong, C. T. 2004. The
Ghoshal, S. 2005. Bad management business school "business": Some lessons
theories are destroying good from the U.S. experience. Journal of
management practices. Academy of Management Studies, 41: 1501-1520.
654 Academy of Management Learning & Education December

Pfeffer, J. 2007. A modest proposal: How we Tsui, A. S. 2007. From homogenization to

might change the process and product of pluralism: International management
managerial research. Academy of research in the academy and beyond.
Management Journal, 50(6): 1334—1345. Academy of Management Journal, 50(6): 1353
—1364.
Rajan, R. G. , & Wulf, J. 2004. Are perks
purely managerial excess? NBER Working UNESCO. 2006. Call for Papers: Universities as
Paper No. 10494. Centers of Research & Knowledge Creation -
An Endangered Species?
Rindova, V. P., Williamson, I. 0., Petkova,
A. P., & Sever, J. M. 2005. Being good The UNESCO Forum on Higher Education,
or being known: An empirical examination Research, & Knowledge. Paris.
of the dimensions, antecedents, and Wooldridge, J. M. 2002. Econometric analysis
consequences of organizational of cross-sectional and panel data.
reputation. Academy of Management Cambridge, MA: MIT Press.
Journai, 48(6): 1033-1049.
Wooldridge, J. M. 2003. Introductory
Rubin, R. S. , & Dierdorff, E. C. 2009. How econometrics. Mason, OH: South-Western.
relevant is the MBA? Assessing the
Wooldridge, J. M. 2005. Simple solutions to
alignment of required curricula and
the initial conditions problem in
required managerial competencies.
dynamic, nonlinear panel data models
Academy of Management Learning &
with unobserved heterogeneity. Journal
Education, 8(2).
of Applied Econometrics, 20: 39-54.
Safön, V. 2007. Factors that influence
recruiters' choice of Bschools and their
MBA graduates: Evidence and implications
for B-schools. Academy of Management

Learning & Education, 6: 217—233.

Scott, J. C. 2006. The mission of the university:
Medieval to postmodern transformations. Joumal of
Higher Education, 77(1):

Skapinker, M. 2008. Why business

ignores the business schools.
Financial Times [Asia Edition],
January 8: 11.

Jonathan P. O'Brien is an assistant professor of strategic management at Rensselaer Polytechnic Institute's Lctlly
School of Management & Technology. He received his PhD in strategic management from Purdue University. O'Brien's
research interests include corporate governance, differences across institutional environments, and the strategic
implications of a firms' financial structure.

Paul Louis Drnevich is an assistant professor of strategic management at the University of Alabama.
He received his PhD in strategic management from Purdue University. Drnevich's research interests
include competitive advantage and value creation/appropriation and the effects of the dynamics of
environmental uncertainty on performance, the implications of capabilities and environmental
factors for innovation and performance in entrepreneurial ventures and small business, and the
application of strategic management theory and methods to solving problems in management education.
T. Russell Crook is an assistant professor of management at the University of Tennessee. He received his PhD in
strategic management from Florida State University. Crook's research interests include strategy and
entrepreneurship topics related to why some firms perform better than others.

Craig E. Armstrong is an assistant professor of entrepreneurship at the University of Alabama. He

received his PhD in strategic management from the University of Texas at San Antonio. Armstrong's
research interests include entrepreneurial decision making, bricolage, and resourcefulness;
entrepreneurship education, absorptive capacity, business models, and strategic competition.