The Theory-Setting-Testable Hypothesis Model
research or by careful consideration of the theoretical model.
The operational linkage can be a possible source of hypotheses because if the linkage is not clear, then research to clarify the relationships could be conducted. A second research possibility can arise from the usual assumption in marketing that the relationship between concepts is linear. It is possible that some relationships may be linear only with- in a limited range. The researcher is more likely to fmd an underlying nonlinear relationship if that possibility is inves- tigated. If a nonlinear relationship is a possibility, the asso- ciated operational linkage should reflect that form.
The operational linkage, how the independent variable is related to the dependent variable, should be clearly speci- fied. Cronin and Taylor (1992) did not do so, as illustrated by one of their propositions: "Consumer satisfaction has a significant impact on purchase intentions." Implicitly, their model assumed a positive direct linear relationship between satisfaction and intentions.
The theoretical statement "As extrinsic satisfaction decreases (X), job seeking is likely to increase (Y)" relates concepts, but understanding is deepened if theory can indi- cate why the concepts are expected to be related (Zaltman, Pinson, and Angelmar 1973); the why involves explanation.
Explanation can be treated analytically by considering the content and type. Content refers to the processes that oper- ate so that X is linked to Y. The type of explanation refers to a formal classification of the scientific generalization, for example, deterministic or probabilistic, that provides the philosophical foundation for the explanation (Hunt 1991).
The processes linking X to Y, or the content of an expla- nation, can often be found by searching for key words used in providing an explanation. One may read that "X influ- ences Y," or "X influences Y by the process of..." (Hage 1972).
been found to be significantly related to X in a number of studies"; however, why Y and X could be expected to be related is not discussed. That omission can be found in Parasuraman, Zeithaml, and Berry's (1988, p. 16) study, which quotes four studies in support of their assertion that "service quality, as perceived by consumers, stems from a comparison of what they feel service firms should offer (i.e., from their expectations) with their perceptions of the per- formance of firms providing the services." However, an explanation of why quality is determined by the comparison of expectations to performance is not provided. In a later article, Parasuraman, Zeithaml, and Berry (1994) provide an explanation, as they argue that expectations serve as norms and norms function as standards "against which perfor- mance is compared."
Understanding can be advanced if the theoretical linkage or explanation is developed in some detail, but the ability to do so is limited. First, a complete account would involve an infinite regression. If one explained that the perception of a problem, insufficient extrinsic rewards, was likely to pro- vide motivation to seek an alternative job, then one would need to explain why low extrinsic satisfaction would result in a perception of a problem and so on. Second, current knowledge may limit the depth of explanation. Determining how far to go in terms of explaining in more detail can be approached as follows. A basic goal of science is explana- tion (Hunt 1991). Specifying only the relationship between two variables without an explanation would provide knowl- edge of an empirical regularity, which is of value. Providing an explanation for the relationship would advance knowl- edge (Zaltman, Pinson, and Angelmar 1973). However, a researcher who has just discovered an empirical regularity may not be able to offer an explanation. In summary, if a researcher can develop the linkage in more detail, a signifi- cant contribution to the literature is more likely.
According to Hunt (1991), the heart of an explanation is that given certain conditions, the phenomenon could be expected to occur. In marketing, most of our statements are probabilistic and take the inductive statistical form: "If X, then Y is likely" (Hunt 1991). An explanation combines (1) facts of the situation so that the inductive statistical state- ment can be applied, (2) the statement "If X, then Y is like- ly," and (3) the expectation that Y will be observed. An illus- tration of an explanation for job seeking follows. First, the facts of the situation are that a number of salespeople for a company are satisfied with their sales work, but dissatisfied with their pay. Second, our theoretical statement was that as extrinsic satisfaction decreases, job seeking is likely to increase. Third, the process linking extrinsic satisfaction and job seeking is as follows: As extrinsic satisfaction with a job decreases, the worker is more likely to perceive that a prob- lem exists, and that perception motivates the worker to search for a solution to the problem. A possible solution would be another job that offers more extrinsic rewards— thus the worker is likely to search for a more extrinsically rewarding job. Finally, we would expect to observe that a higher proportion of extrinsically dissatisfied, compared with satisfied, salespeople search for an alternative sales job.
Providing an Explicit Theoretical Statement
A research study should provide a clear, explicit theoretical statement, because it is the foundation for an empirical study. That appears to be a problem in Parasuraman, Zeithaml, and Berry (1988). They test the implicit hypothe- sis that as perceptions exceed expectations (X), service qual- ity (Y) increases. The sentence in Parasuraman, Zeithaml and Berry (1988, p. 17) that appears to serve as a theoretical statement reads, "Perceived service quality is therefore viewed as the degree and discrepancy between consumers' perceptions and expectations." Without a explication of "the degree and discrepancy," it would not be clear how quality was related to perceptions and expectations. A close reading of the entire article would make it possible to understand Parasuraman, Zeithaml, and Berry (1988). A clear theoreti- cal statement was provided by Teas (1993) who wrote that perceived service quality increases as the differences between performance minus expectations increase across attributes.
Using Partial Formalization to Evaluate a Theoretical Structure
An omission in some marketing articles is a treatment of theoretical linkages. Such articles may provide an extensive discussion of prior research that essentially says, "Y has
An important element of a theory is that the set of statements forming the theory must be systematically related. A useful process for checking the interrelationships among state-