The Qualitative Report 2012 Volume 17, Article 52, 1-55 http://www.nova.edu/ssss/QR/QR17/thai.pdf Straussian Grounded-Theory Method: An Illustration Mai Thi Thanh Thai HEC Montréal, Québec, Canada Li Choy Chong University of St. Gallen, Switzerland Narendra M. Agrawal Indian Institute of Management Bangalore, Bangalore, India This paper demonstrates the benefits and application of Straussian Grounded Theory method in conducting research in complex settings where parameters are poorly defined. It provides a detailed illustration on how this method can be used to build an internationalization theory. To be specific, this paper exposes readers to the behind-the-scene work to develop a theory on the internationalization of small and medium-sized enterprises based in transition economies. It describes each step from sampling to coding and then to theory formation, explaining the rationale each step of the way. The readers can therefore see how a theory took shape and develop from raw data to refined theoretical propositions/hypotheses. Key Words: Qualitative Method, Straussian Grounded-theory, SME, Internationalization, Transition Economy. Background Information It has been widely acknowledged that existing theories on firm-based internationalization are unable to explain the internationalization of small and medium- sized enterprises (SMEs) in transition economies (Griffith, Cavusgill, & Xu, 2008; Pisani, 2009). Firstly, the theories were constructed with data from developed economies (Canabala & White III, 2008; Werner, 2002) where the fundamental conditions for internationalization (such as availability of information, market structure, institutional stability, property rights, etc.) greatly differ from those in transition economies (Nee, 1992;; Peng, 2000; Peng & Heath, 1996). Therefore, analysis of the institutional variables of countries in transition is becoming an important component of our understanding of the globalization of business (Pisani 2009). Furthermore, several assumptions underlying the existent theory are far removed from realities in transition economies (see Barney, Wright, & Ketchen, 2001; Benito, Petersen, & Welch, 2009; Buckley & Casson, 2009; Dunning, 2001; Johanson & Vahlne, 2009; Peng, 2001). For example, nontransparent information flows and lack of recourses and experience inhibit SMEs in these countries to acquire extensive knowledge of their domestic markets, let alone international markets. Moreover, managers of such firms are not free to choose the most efficient entry mode but the existent literature tends to assume so (Brouthers & Hennart, 2007). As a result, several theoretical statements (such as “firms internationalize in order to boost their profits and performance”; “firms internationalize in a systematic and planned manner”; “internationalization of firms is 2 The Qualitative Report 2012 guided by its learning curve and follows sequential stages” etc.) do not appear to be applicable to firms in transition economies (Bjorkman & Forsgren, 2000; Masurel & Smit, 2000; Turnbull, 1993; Papadopoulos, 1988). It has been argued that how and why firms internationalize are dependent upon their relationships and external environment (Granovetter, 1985) but we still lack an empirically based theory that shows which of the many possible contextual factors influence the firm’s internationalization process and how. On the other hand, the literature built on data from transition economies and emerging markets has not yet addressed the process of internationalization as a whole although it has touched certain aspects of internationalization such as organization of joint-ventures in transition economies’ turbulent environment (Mainela & Puhakka, 2009), export orientation of returnee entrepreneurs (Filatotchev, Liu, Buck, & Wright, 2009), factors influencing outward foreign direct investment (Yiu, Lau, & Bruton, 2007), drivers of internationalization from developing countries to developed countries (Yamakawa, Peng, & Deeds, 2008), and so on. Nevertheless, these studies strongly suggest that firms in transition economies behave very differently from firms in advanced economies. For these reasons, transition economies present an interesting context that can help us clarify the existing theory and moves it forward. SMEs from these economies not only have to face with the well-known problems typically encountered by SMEs (e.g., lack of managerial and marketing skills, lack of financial resources and so on; (see Huang & Brown, 1999 for a full list), but also are further constrained by the unstable institutional environment at home as well as their lack of international knowledge and experience. For them, doing business abroad is like taking cautious steps into unknown territory rather than as a consequence of a rational choice based on economic analyses (Masurel & Smit 2000). Therefore, internationalization literature in the behavioral school of thought became our important starting point for our quest to understand the nature of the internationalization process of SMEs based in transition economies. We sought to answer two fundamental questions in process research suggested by Pettigrew (1992) and Van de Ven (1992): (a) Why do SMEs based in transition economies choose to internationalize and (b) How do they internationalize? The objective for answering the first question is to identify factors that motivate these firms to internationalize. And the objectives for answering the second question are to uncover their internationalization strategies and to learn about their internationalization strategy development. We found that it was impossible to get a satisfactory answer from the existing theories (most notably, stage model of internationalization (Johanson & Wiedersheim-Paul, 1975), internationalization process model (Johanson & Vahlne, 1977, 2009), innovation-related models (Bilkey & Tesar, 1977; Cavusgil, 1982; Reid, 1981), and entrepreneurial value creation models (Jones & Coviello, 2005; Schweizer, Vahlne, & Johanson, 2010) although they could serve as the stepping stones to explore the phenomenon. As it is not possible to deduct an answer from the literature and pertinent concepts and their relationship are unknown, we needed a research methodology that allowed for the development of conceptual categories and their relationships from raw data (Patton, 2002). Mai Thi Thanh Thai, Li Choy Chong, and Narendra M. Agrawal 3 Research Setting Rationales for Choosing Vietnamese SMEs as a Sampling Frame We chose Vietnamese SMEs as a sampling frame for several reasons. Firstly, Vietnam’s economy and Vietnamese SMEs have been internationalizing at a remarkable speed (GSO, 2007; Kokko & Sjöholm, 2004; UNCTAD, 2006). Secondly, its institutional environment has all the characteristics representative of a transition economy such as distorted information, weak market structures, poorly specified property rights and institutional uncertainty (Cheng, 2009; Economist Intelligence Unit, 2009; Nee, 1992; ONI, 2007; Peng, 2000; Peng & Heath, 1996; Vo, 2004). Thirdly, SMEs account for over 99% of all business establishments in Vietnam and respectively 80.6% and 84.2% of the total number of enterprises participating in import-export activities of this country (ASMED, 2006). Research in Vietnam When we went to the field to collect data, we encountered the same challenges that Napier et al. described in their discussion about doing qualitative research in Vietnam (Napier, Hosley, & Nguyen, 2004). Firstly, research is viewed by academics and practitioners in Vietnam as “solution-oriented work,” so studies aiming at developing and testing theories are not appreciated as they are in western countries. Secondly, in order to gain legitimacy in conducting research, the researchers needed to be senior scholars with strong network within university and/or government organizations. Thirdly, mail surveys often do not work because of the unreliable mail system, strong resistance from research subjects as Vietnamese do not like to do things for strangers particularly without payment, and lack of support for research that does not directly benefit the respondents in the short-run. Fourthly, business managers typically see academics as being too theoretical and removed from practice and thus managers do not believe that academics can contribute anything useful for their business. For these reasons, they do not appreciate research and are reluctant to respond to researchers. Fifthly, practices such as recording and taking notes during interviews are considered unimportant by researchers in Vietnam. Vietnamese culture teaches its people to be modest (Vượng, 2000) and people in Vietnam do not have the freedom of speech as westerners do. As a result, young researchers without affiliation to local university or government must rely on personal network to recruit respondents and be creative in finding measures to protect their respondents. Besides knowing the researchers in person, respondents often request face-to-face communication in responding to the researchers. Research Strategy Rationales for using Grounded Theory As one of the available strategies in qualitative research classified by Patton (2002), Grounded Theory was found to be the most appropriate. This scientific method is concerned with the detection and explanation of social phenomena (Haig, 1995) by 4 The Qualitative Report 2012 answering socially purposeful questions of what is happening and why (Douglas, 2004) – It has been well argued that Grounded Theory is particularly suitable when “the topic of interest has been relatively ignored in the literature or has been given only superficial attention” (Goulding, 2002, p. 55). Such is the case with research on the internationalization of SME based in transition economies. The method well serves the purpose of this research project, which is to generate a theory that can explain a process (Patton, 2002). And very importantly, Grounded Theory enables researchers to generate theories that “(1) enable an explanation of behavior, (2) are useful in advancing a theory, (3) are applicable in practice, (4) provide a perspective on behavior, (5) guide and provide a style for research on particular areas of behavior, and (6) provide clear enough categories and hypotheses that crucial ones can be verified in present and future research” (Goulding, 2002, p. 43). Straussian vs. Glaserian Methods Since its introduction in 1967, Grounded Theory has been progressively developed as “a problem-solving endeavor concerned with understanding action from the perspective of human agents” (Haig, 1995, p. 281). The evolution of the Grounded Theory method has led to the formation of two camps, each subtly distinguished by its own ideographic procedures: the Glaserian School and the Straussian School. Although academic research seldom states the differences between the two schools (Locke, 1996; Goulding, 2002), we had to make a choice between the two approaches because they dictate very different practices The Glaserian School stresses the interpretive, contextual and emergent nature of theory development. Glaser (1992) states “The research question in a Grounded Theory study is not a statement that identifies the phenomenon to be studied. The problem emerges and questions regarding the problem emerge by which to guide theoretical sampling. Out of open-coding, collection by theoretical sampling, and analyzing by constant comparison emerge a focus for the research” (p. 25). Therefore, it’s very strong dicta are: (a) “do not do a literature review in the substantive area and related areas where the research is to be done” and (b) “when the Grounded Theory is nearly completed during sorting and writing up, then the literature search in the substantive area can be accomplished and woven into the theory as more data for constant comparison” (Glaser, 1998, p. 67). Given that our research was led by research questions, the Glaserian Grounded Theory approach was deemed inappropriate. Moreover, our prior knowledge in the area (from our previous training and research) would make this demand of the Glaserian approach (going into the field without any idea of what to look for) unrealistic. In addition, this research project was subject to time and resource constraints, and hence it had to be completed within a pre-determined time period. The Straussian School, on the other hand, emphasizes highly complex and systematic coding techniques and permits a preliminary literature study to identify research problems and the areas in which to look for data. According to Straussian guidelines, “…the researcher does not begin a project with a preconceived theory in mind” but rather “ he/she begins with an area of study and allows the theory to emerge from the data” (Strauss & Corbin, 1998, p. 12). Therefore, we considered Straussian Grounded Theory to be more appropriate than the Glaserian one for our research project. Mai Thi Thanh Thai, Li Choy Chong, and Narendra M. Agrawal 5 In this project, we had to take special care to “…avoid imposing concepts that reflect his/her own epistemological predilections other than those emerging from interaction with the study site, its participants and subsequent data” (Douglas, 2004, p. 60). Straussian Research Process Our research design followed the contemporary guidelines set forth by Straussian theorists (Figure 1). In brief, this mandates a recursive process of data collection, data coding, comparative analysis, and theoretical sampling until theoretical saturation (Glaser & Strauss, 1967; Goulding, 2002; Locke, 1996; Strauss & Corbin, 1998). It requires “not only that data and theory be constantly compared and contrasted during data collection and analysis but also that the materializing theory drives ongoing data collection” (Locke, 1996, p. 240) Figure 1. Grounded Theory’s Recursive Analytic Operations (Locke, 1996, p. 240) Generate Compare Collect data Theoretical conceptual data sampling categories observations Theoretical Theoretical sampling sampling (existing data set) Generate theoretical statements With an initial set of data, we conducted open-coding to generate conceptual categories and determine how the categories vary along different identified dimensions. Next, we carried out axial-coding (i.e., comparison of data observations) to find the relationships that existed between the different categories of conceptual properties and dimensions. Then we did selective-coding to integrate the different categories in order to build a theory with theoretical statements. To validate the theoretical statements (propositions or hypotheses) of the emerging theory, we collected new data with theoretical sampling logic and went through the three coding processes again. The task was recursive and completed only when we reached theoretical saturation, which is the point where the theory is deemed refined and thoroughly validated (because additional new data will not add any new idea or understanding to the question at hand). At the beginning of the data gathering process, we did not fix the number of cases, but continuously added more cases until we reached the theoretical “saturation point” (Glaser & Strauss, 1967; Tashakkori & Teddlie, 1998). After the pilot case had been analyzed, subsequent cases were selected based on the analysis of the previous 6 The Qualitative Report 2012 cases. Following Strauss and Corbin’s guideline (1998), we did not try to control variables or look for representativeness or distribution of population; rather, we looked for how concepts vary dimensionally along their conceptual properties. In other words, new evidence found at each stage was used to modify or confirm the emerging theory, which then pointed to an appropriate choice of theoretical possibilities (instances) to investigate in the next phase (Denscombe, 1998). Our sampling of Vietnamese SMEs stopped after 35 cases, when we could not find any new evidence which could inform or underpin the development of theoretical points of the emerging theory. It should be noted that all cases were checked against the database provided by the Vietnamese government’s Agency for SME Development to make sure that they qualified as SMEs according to the official definition developed by the Vietnamese government. Data Collection Sampling. As the research process is a process of discovery rather than hypothesis testing (Denscombe, 1998), sampling for Grounded Theory analysis must not obey the principles of statistical sampling but rather the principles of theoretical sampling as defined by Strauss and Corbin (Strauss & Corbin, 1998). At the outset of the sampling process, we reasoned that we needed to select an SME that had been involved in as many modes as possible (covering inward, outward, and cooperative operations) in order to build the initial structure for the theory. Of the possible candidates, the pilot case was selected by convenience, access, and geographic proximity (Yin, 2003). The founder of the first case is actually a long-time friend of the first author of this paper, and he is now a professor of entrepreneurship at a university in the United States. Their relationship and his understanding of the literature allowed for in-depth interviews in which we discussed the phenomenon from both theoretical and empirical perspectives. The considerable insight from this pilot study enabled us to design better- structured questions to ask subsequent cases. At the company’s requests, we had to send our questions one month in advance (Appendix 1) and we were subject to time contraints so we could not wait to complete analyzing one case before sending questions to the next. Furthermore, it is unlikely that the direction and theoretical structure uncovered from this pilot case will be largely irrelevant to other cases. Therefore, it made sense to send out standard questions for the informants to understand our research interests in discussing with them. We want to emphasize that this set of questions was only to start our interviews and it was by no mean the boundary of our discussion. Our informants were free to explore the phenomenon in all directions that came up to their minds. Our additional questions during the interviews stemmed from their response as well as from findings from previous cases – the purpose being to find out if we should confirm, modify or reject our theoretical statements (hypotheses or propositions) and to decide on what to do next. For example, when the informants of case 2 talked about how internationalizing their firm gave them the satisfactory feeling of being seen bold and innovative even when they were uncertain if internationalization of their firms made any economic sense. This idea resonated what we found in case 1 that personal desire could be an important driver of firm’s internationalization. To validate this proposition, we added the question “what is the role of your personal desire in your firm’s international strategy development?” to all of our subsequent interviews. Another example is that we Mai Thi Thanh Thai, Li Choy Chong, and Narendra M. Agrawal 7 followed these informants’ idea of “being innovative” and made constant comparison to innovation-related models (Bilkey & Tesar, 1977; Cavusgil, 1982; Reid, 1981) because the idea appeared to validate these models. At the end of our project, we were able to confirm their proposition that “individual decision-maker(s), not the firm’s structural design, is the most critical factor in the internationalization process”. However, we had to reject their proposition that “internationalization process is similar to step-by-step sequence of adopting a new technology, extending from the point of no involvement to the point when international activities are regarded as an ordinary and accepted part of the firm’s activities” After the pilot case had been analyzed, subsequent cases were selected based on the analysis of the previous cases. Following Strauss and Corbin’s guideline (1998), we did not try to control variables or look for representativeness or distribution of population; rather, we looked for how concepts vary dimensionally along their different properties. In other words, new evidence found at each stage was used to modify or confirm the emerging theory, which then pointed to an appropriate choice of instances in the next phase (Denscombe 1998). The following example is illustrative of the process: Firm C1 internationalized immediately after its inception and its founder considered market demand to be the strongest driver; thus, we hypothesized that demand factors could influence the firm’s internationalization timing. To validate this proposition, we analyzed two more firms which started internationalizing at relatively the same time and operated in the same industry (to make sure the external conditions were the same); even though, one of them internationalized immediately while the other internationalized ten years after its inception. Readers interested in how the rest of the cases were selected can consult Appendix 2. Data Sources This research relied on multiple sources of evidence (Table 1). Of these sources, qualitative, in-depth, semi-structured interviews was considered to be the most important data source because, as they provided us with richer and deeper insights into the complex phenomena under investigation by answering why and how research questions (Easterby- Smith, Thorpe, & Lowe, 1994; Eisenhardt, 1989; Perren & Ram, 2004). Table 1. Sources of Evidence and Their Strengths (Adapted from Yin, 2003) Source of evidence Strengths Interviews targeted - focuses directly on the topic insightful - provides perceived causal inferences Documentation stable - can be reviewed repeatedly unobtrusive - not created as a result of the case study exact - contains exact names, references, and details of an event broad coverage - long span of time, many events, and many settings Archival records [same as above for documentation] precise and quantitative 8 The Qualitative Report 2012 Interviews. In our project, we worked in a team of three. The first author is a Vietnamese who grew up in Vietnam, studied cultures, and had extensive work experience as a business woman and then a reporter for a foreign newspaper in Vietnam before going to the US for her MBA and Switzerland for her doctoral studies. The second author is a Singaporean professor who directed a Switzerland-based research center on Asia. And the third author is an Indian professor who was based in India and did research on organizational behaviors. Both of these authors were her dissertation advisors. Given the fact that gaining access to informants in Vietnam requires a strong local network, the first author was responsible for conducting interviews. The journalism techniques that she had learned during her work for the newspaper became a strong asset. As she is a member of the same cultural and institutional context as the respondents, they could help us understand the dynamics behind the respondents’ decisions (Peterson, 2004). For example, when Vietnamese say “eco-tourism”, they mean “completely creating a new ecosystem for the purpose of tourism” (such as making artificial cascades, fish ponds, etc.). However, westerners will understand “eco-tourism” as “tourism to places having unspoiled natural resources, with minimal impact on the environment being a primary concern.” Therefore, it was necessary to have in our team a person who could understand exactly what the informants meant and acted as our context translator during the data collection and analysis phases. Indeed, being able to interview in the language spoken by the interviewees allowed the informants to express their thoughts with all subtlety as they desire and this permitted us to limit language biases that can be problematic in cross-cultural interviewing (Marschan-Piekkari & Reis, 2004). Our interviewees were managers who were deeply involved with key decision- making processes in their respective firms in regard to making and implementing their internationalization strategies. Since these managers are clearly the most knowledgeable sources of information about their international activities (Shook, Ketchen, & Cycyota, 2003), we believed the interviews with them were the most important source for insightful first-hand information. Depending on the nature of decision making system of each firm and also the willingness of potential informants, we interviewed one or more people in each firm. More specifically, if firm’s decisions were made in group, we had group interviews. On the same token, if firm’s decisions were made by individuals, we only interviewed the individual decision makers. To learn about the decision making structure of the firm, we usually learned from the secretary and/or personal assistant of the managing directors of the firm as well as our reference sources before making official requests for studying the firms. It should be noted that our case selection followed theoretical sampling logics so we generally had certain ideas of the firms before studying them in depth. We agreed to keep their company identity strictly confidential in order to get their permission to record the interviews with them. At their request, we sent them preliminary interview questions three weeks in advance so that they had time to reflect on their answers. During the interviews, we let the key respondents talk freely about their international development, and we tried not to influence their answers at that juncture. After the respondents had completed their recollections, we asked them another set of scripted questions concerning topics not yet covered. Some were conducted online via teleconferencing programs such as Yahoo and Skype, and some were conducted via Mai Thi Thanh Thai, Li Choy Chong, and Narendra M. Agrawal 9 email, and still others were conducted via telephone. The choice of interview medium was determined by the respondents. Each informant was interviewed twice (average of two and a half hours the first time and one hour the second time) over the course of 18 months, from October 2005 to March 2007. The first round of interviews was to generate new data. The second round of interviews was to verify the explanatory power of the theory we developed. In this second wave, we presented our model and asked the informants if our model correctly depicted their internationalization process and the rationales behind it. Our discussion was focused on their feedback on the model in order to refine it and eventually validate our findings. Documentation and archival records. After making clear to them the purpose of this study, we were allowed to study some of their company documents a month before the actual interviews so that we could pose more informed questions. Since the major reason for the inapplicability of existing theories in the context of transition economies is the institutional environment, it is necessary to understand the impact of the environment in which these firms operate. Therefore, we analyzed local and international newspapers, statistics reports by the World Bank, the United Nations, and Vietnam’s General Statistics Office, and industry reports indexed in Factiva databases. The aim was to understand the environment in which these firms operated and to evaluate the accuracy of decision makers’ perception of their market conditions, industry structures, and business environment in Vietnam as well as in their host countries. Unlike our informants, we had access to both Vietnamese sources and international sources because our research team is international. This allowed us to check if the information available to our informants was distorted and if they missed anything important. Opportunity recognition and strategy development are dependent on managerial perception and cognition which in turn are dependent on information acquired. Like in other countries in transition, information in Vietnam is subject to strong censorship. Hence, comparing the public knowledge against the knowledge that mangers had acquired for their decision making was an important task. Throughout this research project, we continually conducted a literature review for background materials against which to compare the findings from the data we had collected. By reviewing relevant literature, we were able to establish a working knowledge of the topic being investigated, thus enhancing theoretical sensitivity. However, we did not use this literature review to compose any hypotheses but to do 9 tasks listed in table 2 (Strauss & Corbin, 1998, pp. 49-52). 10 The Qualitative Report 2012 Table 2. Illustration of the Use of Literature Task Illustration (1) compare properties and We constantly compared concepts defined in resource-based literature dimensions of concepts derived (e.g. Wernerfelt, 1984; Andersen & Kheam, 1998; Barney, et al., from the literature to our data in 2001; Peng, 2001; Westhead, et al., 2001; Sharma & Erramilli, 2004) order to differentiate and give and organization theory and design literature (e.g. March, et al., 1958; specificity to the emergent Tirole, 1988, Daft, 2004; etc.) with the concepts that emerged from concept; our data on firm’s internal factors and their influence on the internationalization of SMEs (2) identify significant concepts The literature helped name codes during our open-coding process and that are found over and over arrange them in axial-coding. For example, “market structure”, again in the literature and which “product life cycle”, “industry life cycle” etc. are found over and over also appear in the data; again in economic literature an FDI literature in which and these concepts belong to “industry structure and attribute” (see Appendix 4). (3) find clues as to what to look Repetition of codes under the “industry structure and attribute” for in our data and to generate category alerted us that these concepts are important. Therefore, we questions to ask respondents; had to dig further during analysis and asked respondents questions around these concepts to find out if they are actually the same or different from the literature. We finally found out that the informants really talked about perceptions of these factors and their understanding of these concepts is very different from what is defined in the literature. This caused us to make new hypotheses. (4) look for how the properties To examine the concept “international experience”, we looked for and dimensions of documented literature on experience as the process of getting international market concepts and relationships vary knowledge and skills in doing business overseas (e.g. Emden, et al., under a different set of 2005; Johanson & Vahlne, 2003; etc.) and literature on experience as conditions; something that happens to firms that affects how they react (e.g. Newbould, et al., 1978; Reuber & Fischer, 1997; etc) to find out which sense should be taken in our study. (5) check quoted materials from The informants gave us information about the market condition in interviews and field notes as certain countries and explained how it influenced their well as descriptive materials internationalization strategies. Therefore, we read literature that concerning events, actions, covered market conditions of the markets they mentioned. This setting, and actors’ perspectives; allowed us to check the accuracy of the knowledge they acquired about the market. We also read articles about market condition as an internationalization determinant (e.g. Dunning, 1980; Etemad, 2004; Chen, 2008; Thai & Chong, 2008; etc.). This allowed us to compare the type of influence this knowledge had on their international strategy development and the type of influence recorded in the literature. (6) formulate questions that act The literature on strategy formation (e.g. Mintzberg, 1978; Spence, as a stepping-off point before 2003; Olson, et al., 2005; etc.) provided us with a list of conceptual and after interviews; areas to investigate. Thus, it helped us make questions to ask