From the Editors

Enhancing Marketing Theory

in Academic Research
David W. Stewart
University of Southern California

George M. Zinkhan
University of Georgia

Over the past 50 years, the discipline of marketing and In the 50 years that have followed, questions about
the body of marketing knowledge associated with it have buyer and seller behavior, and the factors responsible for
made enormous advances. In 1955, only two well-recog- such behavior, have been addressed by marketing scholars
nized marketing journals existed: Journal of Marketing and with increasingly sophisticated conceptual and method-
Journal of Retailing. The articles published in these two ological tools. At the same time, there have been improve-
publications, with a few exceptions, tended to be highly ments in business practice and important advances in
descriptive and focused on the institutional dimensions of related social sciences such as psychology, economics,
marketing practice. The domains of consumer research and and sociology, as well as advances in research methods,
quantitative marketing that now dominate much of market- including those in data collection, analysis, and statistical
ing scholarship were still embryonic and not yet recognized inference. During much of the past 50 years, marketing
as the disciplined, theory-based approaches to scholarship research and knowledge have advanced by borrowing and
and marketing practice that they have become. adapting theories from other disciplines. More recently,
It was in 1956 that Wendell Smith wrote the first aca- marketing scholars have begun to develop theories more
demic article that suggested that consumers differ on specific to markets and customer behavior and have added
important dimensions. This important article, the first on unique perspectives to those theories borrowed from other
“segmentation,” was followed 1 year later by Pierre disciplines. Marketing scholarship has matured as it has
Martineau’s (1957) important book on “motives that more and more translated empirical findings into theories
make people buy.” Markets and the buyers and sellers of buyer and seller behavior, of markets, and of marketing
within them were recognized as varied, complex, and institutions and practices. These advances in conceptual
worthy of study at the individual and segment level, as development, and the unique issues associated with con-
well as at the level of the aggregate market. Within a few tributing to theory and conceptual thinking in the field of
short years, the Ford Foundation would energize such marketing, are the focus of this editorial comment.
study through its support of John Howard’s Buyer Specifically, we concentrate on marketing theory as it is
Behavior Project and its seminal business education pro- developed and presented in academic journals. Conceptual
ject, which facilitated greater integration of mathematics and theoretical articles are arguably the most difficult
and the behavioral and social sciences into business types of scholarship to craft in a compelling fashion that
scholarship and education. make a significant contribution to the discipline. It is cer-
tainly the case that it is more difficult to get conceptual
articles through the review process. Nevertheless, strong
Journal of the Academy of Marketing Science. conceptual and theoretical articles provide the founda-
Volume 34, No. 4, pages 477-480.
DOI: 10.1177/0092070306291975 tions for subsequent empirical work and serve to integrate
Copyright © 2006 by Academy of Marketing Science. and extend past empirical work. In our view, there is a

Downloaded from jam.sagepub.com at SAGE PUBLICATIONS on December 6, 2012

478 JOURNAL OF THE ACADEMY OF MARKETING SCIENCE FALL 2006

need for more such work in the field of marketing, and the number of products and industries. Thus, there is a real
ability of the field to produce such work is an important challenge for marketing scholars to create generalizable
indicator of its health and maturity as a discipline. We knowledge. Good theory building and theory testing are
would like to see more such work published in Journal of ways to try and achieve such a goal.
the Academy of Marketing Science (JAMS), but not at the Good theory does more than describe; it identities
expense of or as a substitute for strong empirical work. causal structures that provide the basis for forward pre-
Neither do we wish to suggest that JAMS will be an easy diction (not just good fits of historical data) and control
place to publish half-baked conceptual articles. JAMS (at least in some circumstances). It provides explanations
welcomes strong conceptual and theoretical articles, but of marketing phenomena—for example, why customers
these articles will be rigorously reviewed and will need to behave as they do, why different markets evolve in par-
meet a high standard. Nevertheless, we want to encourage ticular ways over time, and why specific institutions and
such work, precisely because it is both difficult to execute practices emerge. With the addition of situational (e.g.,
well and can provide a significant contribution to the field. industry) knowledge), marketing theories may also pro-
In this editorial, we offer some suggestions to guide vide for predictions. While it is the case that the social
the crafting of such work. In doing so, we rely on our sciences, in general, often have difficulties in predicting
experiences as scholars, editors, and review board mem- patterns of human behaviors (Secord 1983), it is a mis-
bers. First, we begin by discussing some attributes related take to dismiss prediction as a necessary characteristic of
to “strong theory,” and then we describe some common good theory. There is a myth that the physical sciences,
problems associated with theory development. We con- the “real” sciences, are somehow better at prediction.
clude with some prescriptions regarding the writing, They are not. They are only better at limiting the tests of
reviewing, and evaluation of conceptual articles that may their predictions to carefully controlled environments
be of use to journal readers, reviewers, and authors. where a test is possible but also bound by the testing
environment. Strong theories are simple and elegant, they
provide rich descriptions, and they provide the basis for
HALLMARKS OF STRONG THEORY testable predictions in controlled circumstances.
Traditionally, there is some disagreement about what
constitutes good marketing theory (cf. Hunt 1990; WHAT THEORY IS NOT
Zinkhan and Hirschheim 1992). Sutton and Staw (1995)
have persuasively argued that theory is not references, Sutton and Staw (1995) argue that descriptions of the-
data, variables, diagrams, or hypotheses. Such elements ory building can leave a reader more, rather than less,
have their place in theory development but none of them, confused about how to write an article that contains
alone or in combination, is a theory. Rather, in our view, strong theory. Instead, they explain that it might be more
strong theory has some of the following characteristics. productive to erect five “warning signs” (or “wrong way”
Strong theory resonates, it shows patterns of intercon- signs). In other words, they describe things that may
nectedness, it provides details about causal mechanisms, appear in an article that are not theory; these include
and it provides answers to the question Why? (Kaplan (a) diagrams and figures, (b) conceptual definitions,
1964). Strong theory captures and succinctly summarizes (c) hypotheses, (d) data, and (e) references. We discuss
knowledge that is generalizable. This goal may appear to each of Sutton and Staw’s warning signs, as they apply to
be a particular challenge in an applied field, like market- marketing thought.
ing, which has its roots in application and practice. If
marketing is to take its rightful place as social science, it Diagrams and Figures
must move beyond these applied roots to focus on more
general questions. Integration of data and empirical find- Many marketing articles include diagrams or figures.
ings across customers, situations, specific markets, and Such devices, on their own, are not theories (Sutton and
specific industries is critical for theory development. The Staw 1995). Instead, the theory builder must provide a
individual market or application may be informative, but rich description in the text so as to explain the underlying
it does not build theory. causal mechanisms and key processes. Creating a set of
Neither does review of numerous individual studies boxes and arrows that show relationships among vari-
of different markets and applications alone result in a ables is not a theory. There must be a compelling logic
theory. Rather, theory emerges first from the integration of about the reasons why the particular set of variables, their
findings and ultimately results from a willingness to step relationships, and the direction of their relationships are
away from the underlying data to take a creative leap that supported by empirical research. Journal authors should
produces insights not to be found in any individual study. describe why this particular model is more compelling
In the end, empirical studies are very often about a limited than any other configuration of variables. Authors should

Downloaded from jam.sagepub.com at SAGE PUBLICATIONS on December 6, 2012

 Stewart, Zinkhan / FROM THE EDITORS 479

develop the logic behind specific, testable predictions there is often a goal to explore research issues that are
about outcomes (even if these predictions can only be relevant to marketing managers. Nonetheless, theory and
tested in highly controlled circumstances). data play a distinct role in behavioral science research
(Kaplan 1964). Data may describe empirical patterns,
Conceptual Definitions whereas theory focuses on causal mechanisms (Sutton
and Staw 1995).
As well described by MacKenzie (2003), defining
focal concepts is a crucial first step for the author of a References
successful manuscript. Conceptual and theoretical arti-
cles are no exception. For example, how can measure- Journal authors provide citations to situate their work
ment be successful if there is only a vague idea of what in the extant literature (Zinkhan, Roth, and Saxton 1992).
exactly is being measured? Nonetheless, a theory is more In other words, authors attempt to anchor their work in
than just a set of clear and concise definitions. A theory prior research, both to justify their approaches and pre-
must go beyond this stage and describe the causal linkage dictions and to make their unique contribution clear.
between concepts. Some social sciences (e.g., sociology) Nonetheless, a listing of references does not constitute a
may be particularly prone to this tendency to “provide a theory. In fact, a single manuscript sentence, followed by
dictionary of a language that possesses no sentences” a listing of seven or eight references, is not an especially
(Weick 1989:517). Marketing has had a tendency to effective way to communicate with journal readers.
engage in such practices. Thus, after more than 20 years Instead, the authors need to discuss prior works in detail,
of research on relationship marketing, there is still an so as to elaborate on the causal logic mechanisms (Sutton
absence of a compelling definition of relationship, and, and Staw 1995).
after 30 or more years of research that “demonstrates” the
importance of involvement in consumer decision making,
involvement is still largely defined in terms of outcome SOME RECOMMENDATIONS
rather than as a more fundamental construct.
Our discussion suggests that the author of a concep-
Hypotheses tual or theoretical article designed for an academic jour-
nal confronts a large number of hurdles. As DiMaggio
Hypotheses represent one kind of “sentence” described (1995) succinctly stated, authors who “disobey Sutton
by Weick (1989). However, a hypothesis is not a theory; and Staw’s injunctions deserve whatever the reviewers
rather, a hypothesis is a derivative of theory. A well- deal them” (p. 396). Editors and reviewers will scrutinize
crafted hypothesis is a statement of what a theory pre- conceptual definitions, references, theories, hypotheses,
dicts and therefore provides for a test of theory. But, a methods, analyses, implications, and more. How is an
hypothesis, on its own, is not a theory. For instance, author to accomplish so much in a rather brief space
hypotheses provide statements about what is expected to (e.g., the length-to-contribution ratio)? One solution is to
occur; they do not necessarily explain why it is expected specialize. For example, JAMS has a long and successful
to occur. Note that some manuscript authors attempt to history of publishing conceptual articles (Zinkhan 2004).
introduce so many hypotheses that it is not feasible to In these cases, the authors can concentrate on the con-
provide an adequate explanation for each and every one ceptual portion of the article and devote only passing
(Sutton and Staw 1995). In this regard, authors would be attention to issues of measurement or analysis. With the
better off with a smaller number of hypotheses that they added space, authors have a chance to sharpen their con-
can conceptually defend through specific links to the pre- ceptual reasoning and provide great depth (e.g., in terms
dictions expected from a given theory. This perspective of discussing causal mechanisms).
also implies that authors (and reviewers and editors) Of course, a journal such as JAMS is interested in both
should be more willing to be satisfied with a journal arti- theory building and theory testing. We urge authors of
cle that tests just one part of theory. As we discuss below, empirical articles to devote considerable energies toward
marketing scholars do not universally accept this kind of describing causal mechanisms and causal structures.
approach. However, such an effort creates a “catch-22” effect for
authors. Specifically, as the authors describe the underly-
Data ing mechanisms (and variables) that underpin their
hypotheses, there is a danger that reviewers (and editors
Data can provide the inspiration for theory (as in induc- and readers) will raise objections. Now, the author has
tive research), but they are not a substitute for theory. introduced variables that are not explicitly measured in
Marketing is an applied discipline, so there is a justifica- the current study. With respect to this issue, we endorse
tion for studying real phenomena from the marketplace; Sutton and Staw’s (1995) recommendation that journals

Downloaded from jam.sagepub.com at SAGE PUBLICATIONS on December 6, 2012

480 JOURNAL OF THE ACADEMY OF MARKETING SCIENCE FALL 2006

“let down their guard just a bit” (p. 381). In other words, fact should not deter such efforts, and journals, including
it may not be necessary for authors to measure every con- JAMS, must work to nurture such scholarship, even in the
cept that is discussed. Instead, journals should be willing face of a very high hurdle. Nonetheless, we believe that
to publish articles that test part of a theory. marketing theorists have made significant progress over the
In many respects, modern marketing theories are rather past 50 years, but, if the discipline is to play an important
complex, and it is unrealistic to expect a definitive test of role in the social sciences and influence management
a specific theory within the confines of a single journal practice, it must accelerate its efforts to create its own
article. Physics does not demand an empirical test of the unique theories of markets and marketing phenomena.
“theory of everything,” but there have also been numerous JAMS is a welcoming and nurturing place for such work.
attempts to offer such theories during the past 100 years.
Efforts to develop such theories must be compelling, and
there have been few real attempts, even in physics. However, REFERENCES
such theoretical developments represent important and DiMaggio, Paul J. 1995. “Comments on ‘What Theory Is Not.’”
creative efforts to integrate vast amounts of empirical data; Administrative Science Quarterly 40 (3): 391-397.
they should not be held to a standard that also requires Hunt, Shelby. 1990. “Truth in Marketing Theory and Research.”
Journal of Marketing 54 (July): 1-15.
empirical proof of all relationships. Nevertheless, such Kaplan, Abraham. 1964. The Conduct of Inquiry. New York: Harper
work should suggest hypotheses that are testable, at least & Row.
in some circumstances. At the same time, a modest empir- MacKenzie, Scott B. 2003. “The Dangers of Poor Construct Conceptu-
alization.” Journal of the Academy of Marketing Science 31 (3):
ical study that tests important predictions of a general the- 323-327.
ory can represent a compelling contribution if it provides Martineau, Pierre. 1957. Motivation in Advertising. New York:
strong support for the theory (and, more important, contra- McGraw-Hill.
Secord, Paul. 1983. “Explanation in the Social Sciences and in Life
dicts any competing theories) or provides unqualified evi- Situations.” Paper presented at the University of Chicago Conference
dence that the theory is incorrect. on Potentialities of Knowledge in the Social Sciences, September,
Philadelphia.
Smith, Wendell R. 1956. “Product Differentiation and Market Segmen-
tation as Alternative Marketing Strategies.” Journal of Marketing
CONCLUSION 21 (1): 3-8.
Sutton, Robert I. and Barry M. Staw. 1995. “What Theory Is Not.”
It may be easier to describe “what not to do” than it is Administrative Science Quarterly 40 (3): 371-384.
to describe “what to do” when writing a conceptual or Weick, Karl E. 1989. “Theory Construction as Disciplined Imagination.”
Academy of Management Review 14:516-531.
theoretical article. In this sense, the five “wrong way” signs Zinkhan, George M. 2004. “Accessing Academic Research Through an
may be of limited value. Nonetheless, it is our experience E-Data Base: Issues of Journal Quality and Knowledge Use.”
that quite a number of marketing manuscripts sometimes Journal of the Academy of Marketing Science 32 (4): 369-370.
——— and Rudy Hirschheim. 1992. “Truth in Marketing Theory
wander down one of these streets where strong theory is and Research: An Alternative Perspective.” Journal of Marketing
rarely present. We realize that some of our recommenda- 56 (April): 8088.
tions are difficult to implement. The reality is that good ———, Martin Roth, and Mary Jane Saxton. 1992. “Knowledge
Development and Scientific Status in Consumer Behavior Research:
theoretical work is far more difficult to do and requires A Social Exchange Perspective.” Journal of Consumer Research
far greater creativity than most empirical research. This 18 (September): 282-291.

Downloaded from jam.sagepub.com at SAGE PUBLICATIONS on December 6, 2012