The Theory-Setting-Testable Hypothesis Model
Readings on major areas of inquiry in marketing, such as consumer behavior or marketing channels, constitute a third source of knowledge from which the doctoral student can draw in developing a dissertation research proposal. A limi- tation of substantive works in terms of learning how to inte- grate theory and research is that the "conceptual essay" does not include research methods and the typically empirical report only presents a discussion of the specific research methods that were used. Usually, philosophy of science ideas remain implicit. As an example, the conceptual and empirical meanings of terms should be congruent, and that can be aided by careful definition at both the conceptual and empirical levels. Different types of definitions are discussed in works on the philosophy of science (Hunt 1991). However, such considerations are typically absent from sub- stantive articles and would not even be considered appropri- ate for most empirical articles.
The divisions between theory, method, and substantive topics noted here are, we believe, reflected in the content of doctoral level courses. In a recent effort to plan a seminar, we contacted a number of schools with well-recognized doctoral programs and requested syllabi for both substanfive and philosophy of science courses. The divisions we discuss seem evident. It is certainly possible that the faculty closes the gaps over a series of seminars. However, a seminar that addresses the issues that our model covers could help devel- op essential integration skills earlier in the doctoral program.
Two lines of evidence—doctoral work and our experience with doctoral students—suggest that our framework could be useful. First, some of the gaps addressed in our model can be found in published research. The need to ground research in theory is well documented, as are repeated calls for mar- keting researchers to do so (Bagozzi 1984; Malhotra 1988). Reiterated calls by tbe editors of the two most prominent journals in marketing to ground research in theory suggests that some researchers are failing to do so (Kinner 1991, 1992; Weitz 1992). Finally, we have served as reviewers for a number of years, and a frequent reason for rejection of manuscripts is that the empirical research lacked clear inte- gration with a theoretical base.
Experience with doctoral students also points to a need for the framework. In our institution, several students have remarked that they had difficulty in seminars taken before exposure to our model because the faculty seemed to assume that they had a basic understanding of "theory." Most of our incoming students have completed a professional master's program, such as a master's of business administration, and students find that our model helps in making the transition in thinking from viewing topics from a professional to an academic perspective. Early in our seminar, many students find that the material is quite difficult, but gain skills over the term.
In summary, the marketing educator must assist doctoral students in learning how to bridge the philosophy of science, empirical methods, and a substanfive topic. Next, tbe theo- ry-setting-test framework is presented as a model for the markefing educator to use as a resource in closing the inte- grafion gap.
Overview of the Theory-Setting-Test Framework for Hypotheses Development
The development of theory-linked hypotheses can be thought of as involving three basic steps: (1) theory and pre- dictions, (2) setting and propositions, and (3) testable hypotheses and measurement (see Table 1). First the researcher reviews the theoretical literature and prior research on the topic and derives a set of predictions to be tested empirically. Second, a setting is envisioned in which the researcher can observe the variables and processes. In the setfing, the theorefical predictions are expressed as spe- cific propositions. Third, operational meastires are obtained for concepts in the propositions. Operational definifions make it possible to specify testable hypothesis that logically relate the theory to empirical observafions. The theory-set- fing-test framework has been used in this arficle to organize the process of hypotheses development. The framework is a logical reconstruction of the research process; not all researchers would use the sequence of steps presented here.
The T Step: Theoretical Statements and Prediction
The heart of a theory is a set of interrelated theorefical state- ments. In its most basic form, a theoretical statement speci- fies that an independent variable (X) is related to a depen- dent variable (Y) (Hunt 1991). The theory seeks to provide an explanafion of why some phenomena are consequences of the independent variable included in the theory. As an ini- fial step, the student is advised to search the literature on a topic of interest for statements linking an independent vari- able to a dependent variable. Once a theoretical statement has been found, it can serve as a basis for a predicfion that the independent variable (X) will be related to the dependent variable (Y). A theory can contain theoretical statements that relate more than two variables; however, for simplicity's sake we have selected a single independent/dependent vari- able form to present the theory-setfing-test framework.
Identification of Theoretical Statements
To illustrate tbe theory to prediction process, assume a stu- dent was interested in voluntary sales force turnover. Prior research has found that as job safisfaction decreased, turnover increased (see Table 1). But the studies failed to disfinguish between salespeople who were dissatisfied with sales as an occupation compared with those who were dis- safisfied with some aspect of their sales job. Theory suggests that job satisfacfion is composed of two major components: (1) intrinsic safisfaction with self-bestowed rewards from the work itself and (2) extrinsic safisfaction with externally bestowed rewards such as pay and other benefits (Churchill, Ford, and Walker 1974). The student reasons that intrinsi- cally satisfied workers who were dissatisfied with extrinsic rewards would seek other employment in sales. On the other hand, intrinsically dissafisfied workers would attempt to change occupadons. The following theorefical statement would be derived: If a worker is intrinsically satisfied, as extrinsic job satisfaction decreases (X), the worker is increasingly likely to seek another job in the same occupa- fion (Y).