Sei sulla pagina 1di 13

Theory Testing Using Case Studies in Business-to-Business Research

Wesley J. Johnston Mark P. Leach Annie H. Liu


The strength of any confirmatory research method depends on two factors. First, the relationship between theory and method, and, second, how the researcher attends to the potential weaknesses of the method. Case research has typically been criticized as lacking objectivity and methodological rigor. As such, case research has been thought to be applicable to exploratory research. By addressing the traditional criticisms of case research, a systematic case methodology is developed that can be useful for testing theory. Central to this confirmatory case method are three elements. First, the reAddress correspondence to Dr. Wesley J. Johnston, Department of Marketing, Georgia State University, University Plaza, Atlanta, GA 30303.

search must begin with hypotheses developed by theory. Second, the research design must be logical and systematic. Third, findings must be independently evaluated. By designing research projects around these aspects, case studies become theory-based, systematic, rigorous, and more objective. As such, case research can provide marketers with one more tool to investigate business-to-business marketing phenomena. 1999 Elsevier Science Inc. All rights reserved.

INTRODUCTION The case study methodology has traditionally been equated with descriptive and exploratory research [1, 2].

Industrial Marketing Management 28, 201213 (1999) 1999 Elsevier Science Inc. All rights reserved. 655 Avenue of the Americas, New York, NY 10010

0019-8501/99/$see front matter PII S0019-8501(98)00040-6

Findings from case research may have more influence on marketing managers than survey results.
In fact, much of the research associated with the case method in marketing has been exploratory in nature [3, 4]. As such, case studies have often been criticized as lacking in rigor, and belonging only to the realm of scientific discoverynot justification [5]. Although previous case studies in marketing have been exploratory in nature, the case study methodology is not limited to exploratory research [6, 7]. Yin [8], for example, argues that the case study method can be used to research questions that are exploratory, confirmatory, or explanatory in nature. As with any research methodology, it is not the qualitative or quantitative nature of case research that makes it exploratory or confirmatory; it is the nature of the research question and the rigorous application of the appropriate methodology. Specifically, for a study to be confirmatory, the research question must be grounded in theory and the methodology must be systematic, rigorous, and objective. The purpose of this article is to discuss the case method as an appropriate tool to conduct confirmatory business-to-business research. A case study methodology is developed by integrating ideas put forth by Yin [8, 9], Hirschman [10], Bonoma [1], and others. This methodology advocates the development of a priori theory-based hypotheses, the use of systematic multiple-case-study research designs, and the use of multiple independent evaluation techniques. Each of the methods three elements have been presented in previous research. For example, Yin [8] advocates the use of theory-based hypotheses and systematic designs. Hirschman [10] argues for the independent evaluation of qualitative research findings. However, only when these elements are integrated does a confirmatory case method emerge. CURRENT METHODS FOR CONDUCTING BUSINESS-TO-BUSINESS RESEARCH In an article dedicated to conducting case research in marketing, Bonoma [1] argues that case studies should become an acceptable form of marketing research. He finds case research to be particularly useful when the phenomenon of interest cannot be easily studied outside its natural setting, and when the phenomenon cannot be readily quantified. As Yin [8] points out, studying phenomena that take place in rich contexts means that there will always be too many variables to consider for the number of observations made. Consequently, the application of standard experimental and survey designs may not be appropriate. Under business-to-business situations where multiple contextual variables influence organizational behavior, and individuals within the organization only see a partial picture of the entire process, survey designs become less appropriate. In particular, the single key informant method has received substantial criticism [1114]. Researchers have addressed difficulties with key informants by developing snowball sampling designs that identify multiple informants [1519]. However, these studies invariably show that multiple informants have different views of the phenomenon [20]. Multiple informant techniques, therefore, address one problem while raising new questions. For example, what is the best way to handle conflicting answers from vari-

WESLEY J. JOHNSTON is a Professor of Business Administration in the Department of Marketing at Georgia State University, Atlanta, Georgia. MARK P. LEACH is an Assistant Professor of Agricultural Economics at Purdue University, West Lafayette, Indiana. ANNIE H. LIU is an Assistant Professor in the Department of Consumer Sciences and Retailing at Purdue University, West Lafayette, Indiana.

202

The strength of case research is its ability to utilize the various sources of evidence and triangulation procedures to demonstrate convergence on one meaning.
ous respondents? Kumar, Stern, and Anderson [20] argue for attaining a consensus by having the informants discuss inconsistent responses in a group. However, group discussion eliminates the confidential nature of the responses. Therefore, seeking a consensus may only be appropriate for research questions that are not of a sensitive or ego-involving nature. Due to the lack of control in real world business organizations, experimental studies (or quasi-experimental studies) have been extremely rare in business-to-business research. Those studies that have been conducted rely on simulated business organizations using student participants [21], or scenarios presented in survey format [22]. The point to this discussion is not to discredit past survey and experimental methods, but to show that every method has faults. These faults may become particularly problematic when conducting business-to-business research. approaches. In consumer research, the vivid, personal, and concrete nature of case study evidence has been found to be more persuasive than statistical findings [23]. Thus, from an organizational change perspective, case research may be more influential than the findings of quantitative research attempts [24]. In business-to-business research, several phenomena may benefit from case research. For example, studies investigating organizational aspects (e.g., how marketing ideas are disseminated through the organization), groups or individuals within an organization (e.g., the buying center or the purchasing agent), the relationships among multiple organizations (e.g., a buyer/seller dyad, or a network of buyers and sellers), or a project or program (e.g., the implementation of TQM, or new marketing strategy) may be of particular interest. Of course, these are not the only business-to-business phenomena that can benefit from case study research, but they illustrate phenomena that take place within rich contexts. Contexts that researchers have little control over and that must be considered in order to fully understand the phenomena. OVERVIEW OF CASE STUDY RESEARCH Bonoma [1] argues that case studies may be most adapted for conducting research in context dependent marketing phenomenon because case studies rely on the use and triangulation of multiple sources of data. The use of multiple data sources is different from other qualitative methods that primarily rely on one method of data collection [25, 26]. Bonoma [1] develops a four-stage process to help guide case research projects in marketing. His four stages are labeled: drift, design, prediction, and disconfirmation. By progressing through these stages, a researcher moves his or her research from an exploratory 203

CASE RESEARCH Case study research consists of a detailed investigation that attempts to provide an analysis of the context and processes involved in the phenomenon under study [8]. No attempt is made to isolate the phenomenon from its context, but instead, the phenomenon is of interest precisely because of its relation to its context. By taking a more systematic and theory-based approach to case study research, case studies may provide a useful, yet underdeveloped, tool to investigate industrial marketing phenomenon. Findings from case research may have more influence on marketing managers than survey results. For example, many firms use case histories (e.g., success and failure examples) for training efforts and developing new sales

If results of case study research are to have a significant impact, they must be compelling and robust.
nature (drift and design), to being more confirmatory (prediction). Finally, the researcher becomes interested in only looking at extreme cases to determine limitations of generalizations (disconfirmation). To date, case research in business-to-business marketing has only addressed the exploratory drift and design stages. For example, Vyas and Woodside [4] employed protocol analysis along with observation and document analysis to examine how business-to-business customers select vendors and what specific choice criteria they tend to use. They conducted multiple case studies (18 cases in 6 firms) to identify choice patterns in buying centers. Hutt, Reingen, and Ronchetto [27] used exploratory single case study techniques and network analysis to assess communication patterns that emerge during the formulation of marketing strategies. Buckles and Ronchetto [28] conducted a single case study integrating network analysis using snowball sampling techniques to examine the purchasing work-flow system within an industrial firm. Buckles and Ronchetto [28] advocate the combination of a case study approach with network analysis to facilitate in-depth inquiry of business-to-business exchange behavior. They suggest this approach provides a richer and more complete explanation of organizational buying decisions than cross-sectional studies. A series of case research conducted in Europe by the International Marketing and Purchasing (IMP) group has developed an interaction framework for business-to-business exchanges [2931]. This interaction approach takes the view that buyers and sellers interact over time and develop a pattern of working relationships. Within these working relationships multiple exchange partners are able to integrate and adapt rules, procedures, and norms to be mutually advantageous. Although all these studies pioneered case method research in the marketing discipline, these examples illustrate how the case approach has primarily been used as an exploratory method. Research conducted by the IMP 204 used multiple cases to confirm findings of preceding cases, however, they were conducted to develop theory rather than to test it. In order to be able to test theory with case studies, more systematic steps need to be taken. These steps include developing a theoretical background, and developing a systematic research design including an evaluation criteria.

A CASE STUDY METHODOLOGY FOR BUSINESS-TO-BUSINESS RESEARCH To use case research to investigate questions of a confirmatory or explanatory nature, three things are needed. First, any case study must begin with theory and the development of research hypotheses. Second, a systematic research design must be followed. Third, an evaluation criteria must be implemented to independently assess potential biases and to ensure the methodological rigor of the case studies. These three elements must be included in any confirmatory case study methodology. However, prior to dealing with any of these elements, one must adopt the idea that case studies should be evaluated as individual studies and not sampling units. Thus, case studies are not generalizable to a sampling universe but instead directly confirm or disconfirm theory [8]. Generalizing from Case Research To use statistical generalizability theory and to conceptualize case studies as sampling units would be inappropriate and unrealistic. Generalizability theory would require establishing an appropriate sampling frame, and conducting a large number of case studies selected from this frame. Given the rich context in which business-tobusiness phenomena take place, the number of variables to consider (and thereby the number of cases required) quickly becomes unmanageable.

Multiple cases allow for replication in multiple settings.


However, the strength of case research is its ability to utilize various sources of evidence and triangulation procedures to demonstrate convergence on one meaning. As such, each case study should be thought of as an independent study that offers great depth and comprehensiveness [6]. Case studies should not be thought of as being a form of data collection, but instead, as whole studies [8]. However, in order to overcome their lack of statistical sampling representativeness, careful selection and replication of cases is imperative. Conceptualizing a case as an individual study, rather than a sampling unit, changes what case or cases are appropriate to investigate. For example, if cases are thought to be sampling units, then the objective is to randomly choose a sample of cases from a population of cases. For various reasons, this is often inappropriate or impossible [1]. By treating each case as a study, the focus shifts to choosing the case or cases that are best suited to investigate the theory. Thus, selecting the case study, or the set of case studies is similar to decisions made by a researcher conducting multiple experiments [8]. Each experiment should complement the others by replicating the findings under various conditions, or by addressing different aspects of the overall theory. The goal is that together the set of studies will provide rich support for the theory. For example, a set of six to eight cases may be specifically selected because they have characteristics that are believed to be applicable to a particular theory. By also selecting a set of cases without these characteristics, the two sets of cases may isolate the contexts in which the theory applies. Yin states, Multiple-case studies should follow a replication, not sampling logic. This means that two or more cases should be included within the same study precisely because the investigator predicts that similar results (replications) will be found. If such replications are indeed found for several cases, you can have more confidence in the overall results. The development of consistent findings, over multiple cases and even multiple studies, can then be considered a very robust finding ([9] p. 34). Similarly, case study results can be very persuasive when multiple cases are used to confirm theorized differences across two or more cases. For example, based on cultural differences, one can theorize how Japanese and American industrial buying behavior may differ. By conducting multiple cases within each of these countries, and if these cases demonstrate the theorized pattern of differences, a researcher can make a powerful argument for the soundness of the theory. Case Studies Begin with Theory This brings up the first of three important elements to case study research. Every case study must begin with theory. Like with all other research methods, it is the degree to which theory and related hypotheses have been developed prior to data collection that allows for the testing of the theory. When research hypotheses do not drive the research, findings can only be thought of as exploratory and/or descriptive. In the marketing literature, several case studies can be found that are descriptive by nature. In fact, many post-modernist researchers argue for an ethnographic approach. This approach seeks to develop a thick description of the phenomenon of interest by deliberately avoiding any preconceived notions or theories [3, 32, 33]. Unlike the ethnographic approach, the case study method developed here requires theory to be consulted prior to data collection. In fact, the research hypotheses guide all of the decisions in the development of the research design. When conducting research using case studies, the research hypotheses may be somewhat different from those used when conducting studies using survey methods. Case studies are not well suited for correlational hypotheses (e.g., as variable x increases, variable y will also increase). Correlational hypotheses are appropriate for use with survey methods, and are therefore, particularly prevalent in current marketing research. Hypotheses suited for case study research include those proposing the existence of a phenomenon (e.g., business exchange 205

Findings from case research should be confirmed by multiple evaluators, and subject to independent evaluation.
takes place in a network of relationships), or the presence or absence of a phenomenon under certain conditions (e.g., there are many service demands by purchasing firms operating in industrial countries, and there are few demands for extra service by firms operating in lessdeveloped countries). As such, hypotheses pertinent to case research are similar to those suited for quasi-experimental methods. Furthermore, case research can investigate hypotheses of a longitudinal nature. By investigating a companys historical documents and conducting in-depth interviews, case research may be better suited for questions addressing causality than survey research. For example, taking a systematic case study approach, an investigation could be designed to incorporate evidence from past documents and interviews to examine the development of channel relationships. Together with other sources of evidence, a rich description of how relationships develop could be pieced together. Survey research addressing this same question would require multiple surveys conducted over the course of several months, if not years. Case Studies Require a Rigorous Research Design The second element of a confirmatory case method requires the development of a research design to rigorously examine the hypotheses. The developed research hypotheses guide the case study research design. When developing the research design, there are three important considerations: (1) one must define the unit of analysis, (2) select the appropriate cases to study, and (3) decide on what data to collect and how to collect it. DEFINING THE UNIT OF ANALYSIS. The first question that needs to be addressed when developing the research design is determining the unit of analysis [9]. For example, when conducting case research on influence patterns in the buying center, what are appropriate 206 cases to study? Should one study an individual, a group, an entire firm, an exchange, or a relationship between firms? In this example, the appropriate unit of analysis is most likely the buying center. On the other hand, if the research addresses how a purchasing agent can become more influential in the buying center, then the unit of analysis may become the individual purchasing agent. Similarly, if the research question pertains to formal and informal relationships between a buying center and a selling firm, the unit of analysis may be the inter-firm relationship. The unit of analysis depends on the research question being addressed. CASE SELECTION. Next, the actual cases to study must be chosen. There are instances when it is only possible to conduct a single-case study design (e.g., a critical case, a rare case, or a unique case) [8]. However, evidence from multiple-case designs are more compelling and make the overall study more robust [34]. Although, single-case studies contribute to a body of knowledge [8, 9], if results of case study research are to have a significant impact, they must be compelling and robust. If case study methodology is to depend on replication logic, then there should be an attempt to fully investigate a phenomenon using multiple cases. Case selection decisions stem from the research hypotheses. Specifically, every case should serve a unique purpose within the overall scope of inquiry. Yin [8] suggests that a researcher should consider selecting multiple cases as one would consider selecting a set of multiple experiments. Each case must be chosen to complement others. Therefore, each case should be carefully selected so that it either predicts similar results or produces contrasting results but for predictable reasons. For example, when conducting five to seven case studies; a few cases (two or three) may be selected because of their similar contexts allowing to directly replicate findings. Other cases would be chosen to pursue other theoretically interesting patterns. Some of these may deal specifically with

confronting rival hypotheses (see Figure 1). The main idea is to choose cases that will provide compelling support for the initial set of propositions. If the findings from any of these cases are in some way contradictory, the initial propositions must be revised and tested with another set of cases, just like what is required when conducting research with other research methods. DATA COLLECTION. Prior to entering the data collection phase, a systematic plan must be developed that stipulates what information is to be sought to fully investigate the research hypotheses and how it is to be obtained. Yin [8] advocates the use of a study protocol which systematically documents all decisions pertaining to the research design, and includes the set of substantive questions reflecting the actual inquiry. The questions presented in the study protocol are not survey or interview questions, but questions that need to be addressed by the case to support the research hypotheses [8]. Thus, each question serves as a reminder to the case investigator as to what information is needed. Each question will, therefore, be accompanied by a set of probable sources of evidence and sample strategies for acquiring that evidence [8]. For each question, in every case, evidence from multiple sources must be collected in order to allow for triangulation. Triangulation refers to the convergence of evidence on one meaning [1, 2, 8]. Thus, it is analogous to a measure of construct validity by showing a case to be internally valid.

One way to collect data from multiple sources is to use more than one data collection method. Case study data collection methods can be either qualitative or quantitative in nature (see Table 1). One possibility is participant observation [10]. This method involves the researcher personally emerging into the phenomenon of interest to get a first-hand account. For example, Price, Arnould, and Tierney [35] investigated the service encounter demands of white water rafting by becoming a customer. In an organizational setting, Burawoy [36] took a job on an assembly line in order to study the creation of workplace practices. Another common technique is to use interviews with a variety of informants. These could range from semistructured open-ended question and answer dialogs, to unstructured interviews. Interviews have been extensively used in past business-to-business research. For example, Johnston and Bonoma [15] developed and investigated the network structure of buying centers through data collected by in-depth personal interviews. Observation is a widely used technique where the researcher is an unobtrusive observer; a fly on the wall. This technique is usually used alongside other methods of data collection. For example, Vyas and Woodside [4] studied the vendor selection process and selection criteria by observing the attendance of meetings between buyers and vendors before conducting in-depth personal interviews that took place directly after these meetings.

FIGURE 1.

Selecting multiple cases to support theory.

207

Table 1 Data Collection Methods for Case Study Development: Strengths and Weaknesses Method Participant observation Strengths Can obtain a firsthand account and an indepth understanding. Provides detailed assessment of interpersonal activities. Weaknesses Not appropriate in many situations. Hard to gain access. Time consuming. Difficulty in assessing objectivity. Potential for Hawthorn effects. Time consuming. Hard to gain access. Potential for Hawthorn effects. Interview questions must be systematically developed. Inaccuracies from poor recall. Potential for interviewees to provide interviewers with the answers they want to hear, or to provide socially acceptable answers. Must be carefully scrutinized for objectivity (what was the purpose of the documents author). May be difficult to access, or access may be deliverately blocked.

Observation

Firsthand account of events and the context of those events.

Interviews

Focuses directly on the case study topic. Provides perceived causal inferences.

Documentary evidence

Produced outside of the research (objectivity). Electronic communication has created numerous new forms of documentation. Precise and consistent. May allow for a review across several years. Can be obtained unobtrusively.

One advantage of case study research is that it can make extensive use of other sources of data within the organization. The use of documentary materials is a valuable source of information. In their investigation of marketing strategy formation, Hutt, Reingen, and Ronchetto [27] used a document analysis to verify information obtained in personal interviews. Among the documents analyzed were internal memos, telephone logs, and selected planning documents. The use of documentation in case study research must be closely scrutinized. Much of documentary evidence has been edited to reflect a desired image. This is especially true of documents meant for the eyes of those outside of the organization. However, documentary evidence is also considered objective because it is generated outside the influence of the research study. With the proliferation of electronic mail (e-mail) and facsimiles as communication devices, and the growing importance of documents presented on web sites for making initial customer contacts, the opportunities for conducting research using documentation appear to be rapidly growing. Although case research is typically thought to include only qualitative evidence, case study data collection techniques can also include questionnaires and other evidence of a quantitative nature. However, questionnaires are usually evaluated alongside more detailed investigations through observation and interviews. In management, Nutt [37] in208

vestigated the role of activism and decision frames on over 300 strategic decisions using in-depth interviews with multiple people involved in each decision. DATA ANALYSIS. The analysis of case research represents assessing whether or not the evidence within each case is internally valid, supportive of the pre-specified hypotheses across the multiple cases, and conclusive. Therefore, data analysis is concerned with the issues of internal validity and generalizability. The issue of internal validity is addressed through an assessment of whether there is appropriate converging evidence to support triangulation. For every case in the multiple-case study, an assessment of each research question must be made. For each research question, only cases that demonstrate adequate triangulation should be evaluated when generalizing across cases. When the evidence from any one case is conflicting or disconfirms the research hypotheses, this evidence must be evaluated and its impact must be assessed. Does the evidence represent a casespecific problem that can be easily rationalized as such? Is it a problem with a particular type of evidence that was poorly conceptualized prior to the data collection? Does the evidence raise doubts in the initial hypotheses and theory? These are the types of questions that need to be raised when determining the severity of conflicting evidence. If the theoretical underpinning of the entire study is found to be in jeopardy, then

the most appropriate action is to revert to an exploratory means of analysis. Thus, analyses should proceed in manner that will help to further develop theory. The issue of generalizability in case research can be addressed in two ways. First, an interpretation and analysis can be conducted to determine whether the multiplecase study confirms the initial research hypotheses. Then, an independent evaluation criteria should be employed to further examine the results and interpretations. The fundamental goal of conducting multiple-case studies is to assemble the theory-suporting evidence from each case. Mulitple cases allow for replication in multiple settings, the investigation of theorized differences across cases, and the possible refutation of rival hypotheses. Thus, an overall assessment must be made to determine if the data across the cases provides sufficient evidence to support the initial theory. Because an overall assessment may lead to unexpected or unusual results, the researcher must be alert to the possible presence of dis-confirming data and the existence of paradoxes and contrasts. Some researchers may argue that the findings of case study research should never be thought of as confirming theory because the analysis is interpretive and subjective. In order to address these criticisms, case researchers should use multiple forms of independent evaluation. Independent Evaluation Like any other methodology, case studies are subject to potential researcher biases. By following systematic research design procedures aimed at confirming a prespecified set of hypotheses, the potential for researcher bias can be decreased. However, decisions that can affect the outcome of the study have to be made at all stages of the research design; from the selection of the cases to study, to the evidence to look for and the means to collect this evidence. Hirschman [10], describing her humanistic inquiry methodology, stresses the importance of assessing the quality of qualitative research. Hirschman provides a set of criteria to evaluate a form of qualitative research she refers to as humanistic inquiry. This criteria includes four concepts: transferability, credibility, dependability, and confirmability. These concepts are thought to be analogous to the concepts of external validity, internal validity, reliability, and objectivity. Several of these concepts have already been addressed in the above discussion of the research design. For exam-

ple, transferability refers to the generalizability (i.e., external validity) of particular findings and is the main reason why multiple-case studies are advocated. Similarly, credibility and dependability refer to the internal validity and reliability of the findings that are addressed in the case method by requiring multiple sources of evidence to triangulate on the same finding. However, Hirschman considers other elements that have not yet been discussed but are important to ensure the validity of case studies. Confirmability specifically addresses the issue of research bias in data analysis and interpretation. In other words, findings are expected to be supportable from the data, be non-prejudiced, and be non-judgmental. However, objectivity can be a problem at any stage of the research process. Hirschman [10] suggests subjecting all study findings to an outside auditor to evaluate whether findings appear to be logical and free from prejudice. In conducting quality case study research, both the study design and the findings should be independently audited to asses areas of potential researcher bias. In particular, assessing the triangulation of evidence within each case, and the assesments of across-case theory confirmation should be conducted by multiple judges. Many accepted data collection methods used in marketing utilize multiple judges to interpret qualitative data (e.g., content analysis of advertising, coding pencil and paper free-response thought listing). Using multiple judges allows for an analytical investigation of interjudge reliability. There are various ways of assessing inter-judge reliability. For example, with more than two judges, Rust and Cooil [38] have recently developed a measure assessing inter-judge reliability that may be more appropriate than alpha. The advantage of interjudge reliability measures is that the objectivity in the interpretation is quantified. Therefore, it can be easily assessed. A second form of evaluation that is discussed by Hirschman [10], but can be adapted to aid case research is credibility. Credibility is the assessment of how well the reality constructed by the researcher is credible to those being studied. To determine credibility the final interpretation is subjected to the scrutiny of those individuals upon whom it is based. In case research, subjecting findings to key individuals in the organization to seek their responses is useful for two reasons. First, misunderstandings may have occurred that may have wrongly affected the results. By going back to reconfirm the findings, these misunderstandings may be identified. Second, by receiving support for the 209

findings by key informants, the results may gain a higher degree of validity. This has been an effective tool for more quantitative results as well. In their study investigating a new quantitative technique to assess businessto-business sales-performance (i.e., data envelopment analysis), Boles, Donthu, and Lohtia [39] took their results back to sales managers for an independent evaluation. Sales managers confirmed the results and, thereby, provided support for the analytic technique. Summary Several researchers have provided guidelines for conducting rigorous case research [8, 9, 34, 40]. Based on this earlier work, a methodology appropriate for business-to-business researchers is developed. This method consists of three elements. First, case research must begin with the development of theory and research hypotheses. Second, a systematic multi-case research design must be developed which is guided by the research hypotheses. Third, the findings from the research should be confirmed by multiple evaluators, and be subject to independent evaluation. Interestingly, these three elements are identical to any other confirmatory research methodology that subscribes to the scientific method. FUTURE RESEARCH Applications of case study research using methods similar to those developed here, have the potential to benefit many areas of business-to-business research. Two areas that may be particularly relevant are the study of network systems, and international business-to-business marketing. Network Systems As the focus of business-to-business markets evolve from the dyadic relations between exchange partners to business networks and systems [15, 41], so must the methods of conducting business-to-business research evolve. Past literature called for a system paradigm [7, 42, 43], and network analysis [40, 44] to better understand business-to-business dynamics. However, most of the business studies investigating these interactions, communication patterns, or influence processes have relied on traditional quantitative research methods such as surveys or personal interviews [15, 17]. Only a limited number of business studies employed qualitative inquiry [4, 27, 28]. 210

Research by the IMP group [2931] has systematically investigated the dependency patterns, and the adaptation of partnerships among firms and their inter-organizational contact persons over time. Not only do these studies represent the most systematic set of case study research undertaken to date, but they bring attention to an important area in business-to-business research that may benefit from more confirmatory case study methods. The interaction approach views the relationship, rather than the individual transaction, as a unit of analysis [31]. It focuses on partnering as a means to exploit resources and achieve mutual goals. Organizational goals focus on the development of desirable bonding and establishing a desirable network position. As such, further study of the interaction approach is likely to benefit from combining the case research with snowball sampling techniques and network analysis. International Business-to-Business Research Studying business-to-business issues across countries is very challenging. Studying them using traditional methods is particularly difficult. For example, when conducting a survey in multiple countries, if these countries speak different languages, surveys must be translated and back-translated showing a high degree of validity [45]. Findings from survey studies assessing differences among multiple countries are highly suspect because across-country differences are confounded by a litany of phenomena. For example, people in countries that are characterized by an individualist culture generally have a strong sense of self-identity, and consistent behavior reinforces this self-concept [46, 47]. Therefore, individualists have been found to provide more consistent and truthful responses under all situations [48]. Thus, when comparing a country with an individualist culture to a country that is not characterized by individualism (e.g., a collective culture), survey differences may only reflect different survey taking techniques. Perhaps more importantly, the same phenomenon (e.g., a quality business relationship) may be represented in two countries in vastly different ways. As such, two questions are of interest to international marketing researchers. First, how is each business-to-business phenomenon manifested within each country. Second, regardless of how the phenomenon is manifested, does the marketing theory apply? In cross-cultural research this is referred to as examining the emic and etic properties of the phenomenon of interest [48]. Thus, any international

marketing research must take into account context. As such, the case research methodology could likely benefit theory testing in this area. The emerging opportunity to develop international networks in Europe and the firmly established business-tobusiness networks operating in Asia suggest that the interaction approach [30, 31] is applicable in international business-to-business marketing. Studying these international networks will be especially challenging and may only be successful through the use of case research. Additional Areas of Future Inquiry Case research may also be useful in studying the new product development process [4951]. The new product development process is typically characterized as a long, multi-person, complex process. By selecting cases of successful and unsuccessful development processes, theories pertaining to this processes could be confirmed or disconfirmed. Another area that may benefit from case research is the study of long-term selling cycles. When investigating sales cycles of several years (e.g., automated systems [52]), survey or experimental research may offer little help. By investigating company procedures and past records, case research may be the only viable way to study these cycles. CONCLUSIONS The objective of this article has been to demonstrate how case research methods can be developed to aid business-to-business researchers. The method presented here draws from the work of Yin [8, 9, 40], who advocates starting from theory and letting the developed theory and research hypotheses drive many of the questions of the research design, and from the work of Hirschman [10] who applies rigorous evaluation techniques to qualitative research. It also draws from the past case research in marketing and its systematic data collection and analysis techniques. It contributes to marketing theory by integrating methodological ideas from case study and qualitative research. Central to the method are three elements. First, theory must guide research decisions. Second, the research design must be systematic and pre-planned. The rigor in research planning must exist at all stages of the research design: from defining the unit of analysis, to selecting the cases to investigate, to collecting and analyzing the evi-

dence. Third, there must be an evaluation criteria that can adequately refute criticisms of subjectivity. Specifically, by subjecting findings from case research to the scrutiny of multiple judges and other forms of independent evaluation, case study results can become very persuasive. The goal of any method is to produce persuasive results confirming a priori theory in a systematic way. Case research still has its limitations. First of all, case research is often criticized for taking extensive time and effort. Compared to other methods, case research is typically more effort intensive. However, the time and effort required depends on the research questions being addressed. Some case study questions can be answered relatively simply. Yin [8] suggests that some case studies can be completed without leaving the library. Furthermore, case study research may be becoming easier to conduct with access to web-sites, e-mail, and other computerized forms of communication. A second limitation of case research is the prejudice against it [1]. This prejudice stems from arguments that case studies are lacking in rigor and reliability, and that they do not address issues of generizability effectively handled by quantitative methods. However, there is nothing about a method per se which makes it weak or strong. The potential strength of any method depends on two factors. First, the relationship between theory and method, and, second, how the researcher attends to the potential weaknesses of the method. In consumer marketing, Sternthal, Tybout, and Calder [53] have distinguished between a confirmatory approach and a comparative approach to theory testing. The confirmatory approach focuses on establishing that the theory is responsible for observed effects. Therefore, manipulation checks and other threats to internal validity are strongly scrutinized. Alternatively, the comparison approach to theory testing focuses on demonstrating that the theory offers a better account of the observed effects than rival theories. By utilizing the three elements of case research presented here, a case study can be designed to effectively address issues pertaining to both of these approaches to theory testing. Through triangulation procedures and the establishment of a rigorous evaluation criteria, issues of construct and internal validity can be addressed. In addition, through the careful selection of cases, rival hypotheses can be investigated. Through the development of more rigorous methods of conducting case research, the case study need not be limited to the realm of discovery but can be effectively used to test theory in complex environments. Business-to211

business phenomena of interest to marketers take place in environments characterized by high levels of complexity. It is here that traditional methods struggle and the case study has the potential to aid research. REFERENCES
1. Bonoma, Thomas V.: Case Research in Marketing: Opportunities, Problems, and a Process. Journal of Marketing Research 22, 199208 (1985). 2. Stake, Robert E.: The Art of Case Study Research. Sage Publications, Thousand Oaks, CA, 1995. 3. Belk, Russell W., Sherry, John F., and Wallendorf, Melanie: A Naturalistic Inquiry into Buyer and Seller Behavior at a Swap Meet. Journal of Consumer Research 14, 449470 (1988). 4. Vyas, Niren, and Woodside, Arch G.: An Inductive Model of Industrial Supplier Choice Processes. Journal of Marketing 48, 3045 (1984). 5. Lutz, Richard: Positivism, Naturalism, and Pluralism in Consumer Research: Paradigm in Paradise, in Advances in Consumer Research, Vol. 16, Thomas K. Srull, ed., Association for Consumer Research, Provo, UT, 1989. 6. Easton, G.: Methodology and Industrial Networks, in Business Marketing: An Interactive and Network Approach, K. Moller and D. T. Wilson, eds., Kluwer Academic Press, Boston, 1995, pp 411492. 7. Johnston, Wesley J., Leach, Mark P., and Liu, Annie H.: Using Case Studies for Theory Testing in Business-to-Business Research: The Development of a More Rigorous Case Study Methodology, in Advances in Business Marketing and Purchasing, Handbook of Industrial Marketing Research, Arch Woodside, ed., in press. 8. Yin, Robert K.: Case Study Research: Design and Methods, 2nd ed. Sage Publications, Thousand Oaks, CA, 1994. 9. Yin, Robert K.: Applications of Case Study Research. Sage Publications, Newbury Park, CA, 1993. 10. Hirschman, Elizabeth C.: Humanistic Inquiry in Marketing Research: Philosophy, Method and Criteria. Journal of Marketing Research 23, 237 249 (1986). 11. Phillips, L. W.: Accessing Measurement Error in Key Informant Reports: A Methodological Note on Organizational Analysis in Marketing. Journal of Marketing Research 18, 395415 (1981). 12. Silk, Alvin J., and Kalwani, Manohar U.: Measuring Influence in Organizational Purchase Decisions. Journal of Marketing Research 19, 165181 (1982). 13. Heide, Jan B., and John, George: The Role of Dependence Balancing in Safeguarding Transaction-Specific Assets in Conventional Channels. Journal of Marketing 52, 2035 (1988). 14. John, George, and Reve, Torger: The Reliability and Validity of Key Informant Data from Dyadic Relationships in Marketing Channels. Journal of Marketing Research 19, 517524 (1982). 15. Johnston, Wesley J., and Bonoma, Thomas V.: The Buying Center: Structure and Interaction Patterns. Journal of Marketing 45, 143156 (1981). 16. Moriarty, Rowland T. Jr., and Bateson, J. E. G.: Exploring Complex Decision Making Units: A New Approach. Journal of Marketing Research 19, 182191 (1982). 17. Ronchetto, John R. Jr., Hutt, Michael D., and Reingen, Peter H.: Embedded Influence Patterns in Organizational Buying Systems. Journal of Marketing 53, 5162 (1989).

18. Spekman, Robert E., and Stern, Louis W.: Environmental Uncertainty and Buying Group Structure: An Empirical Investigation. Journal of Marketing 43, 5464 (1979). 19. Bristor, J. M.: Influence Strategies in Organizational Buying: The Importance of Connections to the Right People in the Right Places. Journal of Business-to-Business Marketing 1(1), 6398 (1993). 20. Kumar, Nirmalya, Stern, Louis W., and Anderson, James C.: Conducting Interorganization Research Using Key Informants. Academy of Management Journal 36, 16331651 (1993). 21. Gundlach, Gregory T., and Cadotte, Ernest R.: Exchange Interdependence and Interfirm Interaction: Research in a Simulated Cannel Setting. Journal of Marketing Research 31, 516532 (1994). 22. Puto, Christopher P., Patton, Wesley E. III, and King, Ronald H.: Risk Handling Strategies in Industrial Vendor Selection Decisions. Journal of Marketing 49, 8998 (1985). 23. Dickson, Peter R.: The Impact of Enriching Case and Statistical Information on Consumer Judgements. Journal of Consumer Research 8, 398406 (1982). 24. Nisbett, Richard E., Borgida, Eugene, Crandall, Rick, and Reed, Harvey: Popular Induction: Information is Not Always Informative, in Cognition and Social Behavior, John S. Carroll and John S. Paine, eds., Lawrence Erlbaum, Hillsdale, NJ, 1976. 25. Wallendorf, Melanie, and Brucks, Merrie: Introspection in Consumer Research: Implementation and Implications. Journal of Consumer Research 20, 339359 (1993). 26. Arnold, Stephen J., and Fischer, Eileen: Hermeneutics and Consumer Research. Journal of Consumer Research 21, 5570 (1994). 27. Hutt, Michael D., Reingen, Peter H., and Ronchetto, John R. Jr.: Tracing Emergent Processes in Marketing Strategy Formation. Journal of Marketing 52, 419 (1988). 28. Buckles, Tom A., and Ronchetto, John R. Jr.: Examining an Industrial Buyers Purchasing Linkages: A Network Model and Analysis of Organizational Buying Workflow. Journal of Business & Industrial Marketing 11(6), 7492 (1996). 29. Hakansson, H.: International Marketing and Purchasing of Industrial Goods: An Interaction Approach. John Wiley, Chichester, 1982. 30. Hallen, Lars, Johanson, Jan, and Seyed-Mohamed, Nazeem: Interfirm Adaptation in Business Relationships. Journal of Marketing 55, 2937 (1991). 31. Turnbull, Peter, Ford, David, and Cunningham, Malcolm: Interaction, Relationships and Networks in Business Markets: An Evolving Perspective. Journal of Business & Industrial Marketing 11(3/4), 4462 (1996). 32. Geertz, C.: Thick Description: Toward an Interpretive Theory of Culture, in The Interpretation of Cultures, C. Geertz, ed., Basic Books, New York, 1973. 33. Glaser, B., and Strauss, A.: The Discovery of Grounded Theory: Strategies for Qualitative Research. Aldine, Chicago, IL, 1967. 34. Herriott, R. E., and Firestone, W. A.: Multisite Qualitative Policy Research: Optimizing Description and Generalizability. Educational Research 12, 1419 (1983). 35. Price, Linda L., Arnould, Eric J., and Tierney, Patrick: Going to Extremes: Managing Service Encounters and Assessing Provider Performance. Journal of Marketing 59, 8397 (1995). 36. Burawoy, M.: Manufacturing Consent. University of Chicago Press, Chicago, IL, 1979. 37. Nutt, Paul C.: Framing Strategic Decisions. Organization Science 9(2), 195216 (1998).

212

38. Rust, Roland, and Cooil, Bruce: Reliability Measures for Qualitative Data: Theory and Implications. Journal of Marketing Research 31, 114 (1994). 39. Boles, Jim S., Donthu, Naveen, and Lohtia, Ritu: Salesperson Evaluation Using Relative Performance Efficiency: The Application of Data Envelopment Analysis. Journal of Personal Selling and Sales Management 15(3), 3149 (1996). 40. Yin, Robert K.: Case Study Research: Design and Methods. Sage Publications, Newbury Park, CA, 1984/1989. 41. Anderson, James C., Hakansson, Hakan, and Johanson, J.: Dyadic Business Relationships Within a Business Network Context. Journal of Marketing 58, 115 (1994). 42. Johnston, Wesley J.: Industrial Buying Behavior: A State of the Art Review, in Review of Marketing, Ben M. Enis and Kenneth J. Roering, eds., American Marketing Association, Chicago, IL, 1981. 43. Bonoma, Thomas V., Zaltman, Gerald, and Johnston, Wesley J.: Industrial Buying Behavior. Marketing Science Institute, Cambridge, MA, 1977. 44. Werner, O., and Campbell, D. T.: Translation, Working through Interpreters, and the Problem of Decentering, in A Handbook of Method on Cultural Anthropology, D. R. Naroll and R. Cohen, eds., American Museum of Natural History, New York, NY, 1970. 45. Iacobucci, Dawn, and Hopkins, Nigel: Modeling Dyadic Interactions

and Networks in Marketing. Journal of Marketing Research 29, 517 (1992). 46. Markus, H. R., and Kitayama, S.: Culture and the Self: Implications for Cognition, Emotion, and Motivation. Psychological Review 98(2), 224 253 (1991). 47. Triandis, Harry C.: The Self and Social Behavior in Differing Cultural Contexts. Psychological Review 96, 506520 (1989). 48. Triandis, Harry C.: Culture and Social Behavior. McGraw Hill, New York, NY, 1994. 49. Dougherty, Deborah: Understanding New Markets for New Products. Strategic Management Journal 11, 5978 (1990). 50. Dougherty, Deborah: Interpretive Barriers to Successful Product Innovation in Large Firms. Organization Science 3, 179202 (1992). 51. Workman, John P. Jr.: Marketings Limited Role in New Product Development in One Computer Systems Firm. Journal of Marketing Research 30, 405421 (1993). 52. Johnston, Wesley J.: Omni Automated Systems, in Cases in Marketing Management: Issues for the 1990s, Charles L. Hinkle, Wesley J. Johnston, and Esther F. Lanigan, eds., Prentice Hall, Englewood Cliffs, NJ, 1992. 53. Sternthal, Brian, Tybout, Alice M., and Calder, Bobby J.: Confirmatory Versus Comparative Approaches to Judging Theory Tests. Journal of Consumer Research 14, 114125 (1987).

213

Potrebbero piacerti anche