Multiple Methods, Communicative Preferences and the Incremental Interview Approach Protocol

Abstract: This article introduces a multiple methods data collection protocol. The protocol has been designed mainly for those unsure of the practical steps needed to insure coherent data-sets with respect to the substantive, theoretical and methodological issues bearing on their research topic. The protocol is based on experience gained during several qualitative studies and over a decade of discussions with several types of researchers. It is also hoped the protocol provokes discussion amongst experienced researchers.

The interplay between purposeful selection and theories of the middle range underpins the rationale for a data collection strategy to be used in qualitative studies where intercultural interaction is obscured by phenomena grounded in two or more cultures. While this protocol will not work with all research topics, it increases data variety and participation regarding many topics, an advantage over survey-based approaches.

The protocol highlights the communicative preferences participants have regarding the types of interaction associated with different research methods, especially interviews and data related to interviews. What remains to be done is the development of sound analysis and analytical linkages from qualitative to quantitative research in mixed-methods designs.

Key words: intercultural communication; social sciences; research methodology; interviewing

This article introduces a multiple methods data collection protocol (here and after protocol). "Multiple methods" should not be confused with mixed methods, the methodological approach that articulates the implications and actual practices of combining quantitative and qualitative data collection and analysis within the same study (CRESWELL, 2002; PUNCH, 2000; YIN, 2003). These terms are elaborated and clarified in Section 3. [1]

This protocol may prove useful in many kinds of studies in which the researcher is considering the use of interview data and types of data complementary to interviews. However, it is especially useful when approaching topics in which intercultural interaction is obscured by phenomena grounded in two or more cultures. The protocol has been articulated for those times when the research questions guiding a particular inquiry may best be addressed by complementary (multiple) methods that are sometimes overlooked in the rush to ‘get interviews'. The protocol highlights linkages between complementary research methods, for example, direct observation, interpretive analysis of thematic or in-depth interviews and biographical correspondence (ADLER & ADLER, 1998; KVALE, 1996; JOHNSON, 2002; PLUMMER, 2001). [2]

This protocol has evolved during studies carried out by my colleagues and I (HOFFMAN, 2007, in press; HOFFMAN, VÄLIMAA & HUUSKO, 2008). The cited studies will be used as illustrative examples for the purposes of this article because several types of data and subsequent analysis have been used including interpretive analysis of thematic interviews, group interviewing and biographical e-mail correspondence. [3]

In addition to empirical studies, I elaborate ideas drawn from over a decade's experience working with masters and doctoral-level students in—and from—several countries. What these students—and their advisors—mainly had in common was a sincere desire to carry out intercultural research on a wide variety of interesting topics in equally interesting research settings. What many lacked, however, was a practical approach that would account for the ontological and epistemological linkages between their research topic, design, analytical strategy, data gathering, analysis and the final write-up of their master's thesis, doctoral dissertation or other types of research projects. [4]

The protocol introduced in this article will not address all of the above linkages regarding all research topics. However, the use of this protocol greatly increases the likelihood of participation regarding many types of topics, which is its' main advantage over survey-based approaches to the same topic. This is because the protocol is very sensitive to the communicative preferences of potential research participants. In simple terms—elaborated in Section 4—potential participants in many types of research situations normally have clear preferences regarding the type of interaction associated with different types of research methods. This protocol explicitly frames the relationship between these methods. Because communicative preferences are quite strong, I make a case for not only paying careful attention to communicative preferences, but tapping into a methodological synergy which seems obvious in hindsight, but which few researchers take advantage of. [5]

In this article, my general approach to research will be staked out, in order that the reader can place the development of this protocol in context. Following this, the use of a clear analytical strategy is underlined in relation to the options that exist for all researchers. I then elaborate the idea of communicative preferences and explain my rationale for developing this protocol. Specifically, I highlight linkages between the nature of interaction associated with different research methods to the specific communicative preferences of potential research participants. This is illustrated by an example protocol from an actual study in progress. The example protocol explicitly links several process features of qualitative research. Finally, some challenges regarding this approach are discussed with particular reference to the features of the qualitative research process that the protocol draws attention to. [6]

As an academic researcher, I have never strayed far from higher education settings. I am especially interested in research topics that underline the power of cultural assumption, manifested in social structure. While both cultural assumptions and social structure are outside the scope of this article, an example would be differences in the way that a particular discipline, like sociology, or interdisciplinary field of studies, like, intercultural communication, are practiced and regarded within particular societies, for example Finland, Canada and France. While many researchers focus more on culture or structure, a focus on the relationship between cultural assumption and social structure illuminates the interesting circumstances when individuals or groups of individuals encounter each other within a very powerful social structure like a higher education system (ARCHER, 1995; BOURDIEU, 1988). This is because higher education—as a social and cultural institution—is capable of both reproducing social relationships, as well as transforming them (BRENNAN, 2002). However, because higher education has been studied a lot (like many research settings) saying something that is original, significant—and interesting—about what is happening in 21st century higher education is methodologically challenging. [7]

For scholars focusing on face-to-face interaction of persons socialized in different cultures, the relevance of sociologically-driven approaches to intercultural communication is not always apparent. However, sociological relevance comes from the thorny observation that it's quite easy to analytically isolate social settings where equity is assumed, but cultural groups present in society are missing—often systematically (HOFFMAN, 2007). Therefore, a focus on face-to-face interaction is irrelevant in many settings. Because an exclusive preoccupation on social interaction—as a research focus—systematically avoids problematic power relations associated with social inequities in many settings, sociologically-driven intercultural communication inquiry is an important stand of intercultural communication research. This type of inquiry forms the critical basis that may lead—in the best of circumstances—to future social interaction amongst groups that know little about the circumstances of the other. [8]

Amongst my close colleagues doing in research in the partially overlapping interdisciplinary fields of international comparative higher education and international migration and ethnic relations and intercultural communication, it is easy to identify several types of differences related to research. The most important and interesting of these differences—to me—is not necessarily the topics my colleagues study, although these are interesting in and of themselves. More interesting are the paradigmatic assumptions that consistently guide my colleagues and I over time. By paradigmatic assumptions, I mean the coherent sets of ideas that clearly signal the type of knowledge different researchers regard as important, the best general approaches to use with regard to that type of knowledge and the procedures for analyzing and presenting data consistent with these assumptions (KUHN, 1996; CRESSWELL, 2002). [9]

The reasons why different approaches to research are interesting are because research is evaluated—in part—with regard to established, highly recognizable conventions associated with particular analytical strategies or traditions (DENZIN & LINCOLN, 1998; CRESSWELL, 2002). The main obstacle many researchers struggle with is coherence between topic, general approach, analytical strategy, data analysis and write-up. This coherence distinguishes high quality research from everything that is not. [10]

It is the search for this coherence; in my own work, the work of my students and collaborators—that is equally interesting as the topics we all focus on. This is because, without this coherence, our studies may not be of much use to the audience(s) that would most benefit from our results. [11]

Following SEALE (2004)—as well as one of the general premises of the Special Issue in which this article appears—I believe it is a good idea to be fairly clear about important underlying features of one's own research work, when given a chance. This is because firstly, these ideas are not discussed in many types of publications, simply because of a lack of space. It is also because, frankly, many readers and audiences are not interested in these types of issues. Mainly, most audiences are interested in the results of a study and possibly how those results were arrived at. But most importantly, when it comes to qualitative methods, clarity of one's approach to research makes it possible to accomplish our goals of research, specifically producing good descriptions, valid interpretations, robust explanations or connecting any of these to evaluations or other types of interventions (HART, 1998). With this in mind, the following are three key assumptions that guide my research practice. This is not an exhaustive catalog, rather a quick glimpse that will hopefully allow the reader to place the protocol presented, in context. [12]

My allowances for a very wide methodological universe should not be confused with "anything goes." More precisely: "Almost anything" can be researched well—in a variety of different ways. Ultimately, it is methodological clarity in conjunction with a very interesting topic that can catch and hold the attention of our audiences. [14]

The term analytical strategy is used here purposefully, as the author has encountered many masters-level and doctoral students who spend a great deal of time thinking and reading a lot about excellent research topics, yet do not spend even a fraction of the time reading books or articles about the analytical approach they will use on that same topic. When thinking about HART's (2001) highly convincing approach to doing a literature search, ignoring the methodological literature bearing on a research topic places any subsequent efforts in considerable doubt. Even though most students have previously read about the elements of research, many mix up general approaches to research, like qualitative, quantitative or mixed-methods; methodological strategies, like the case study or grounded theory and research methods, like group interviews or surveys. It is not uncommon to hear these terms used interchangeably, as if they meant the same thing. Unqualified or vague statements like "I'm going to use interviews" or "I'm going to do the qualitative approach" are red flags that the practical linkages between what CRESWELL (2002) terms framework elements might need some attention. [15]

One might assume that strong national, institutional, disciplinary or even departmental traditions regarding research design, comparative methodology and instruction in the use of research methods, obviates the types of uncertainties mentioned above. However, because of global trends in structural change involving higher education, international migration, academic mobility patterns, interdisciplinary trends and demographic changes in the populations of many countries, these types of assumptions may have little basis in practice, in many locations (MARGINSON & VAN DER WENDE, 2007; HOFFMAN, in press). [16]

In a very practical text on research design, CRESWELL (2002) identifies three general approaches to research; qualitative, quantitative and mixed methods approaches. Each of these general approaches, in turn, has very clear options with regard to analytical strategies that are often used within those approaches. Analytical strategy, in this sense is used for the same reason DENZIN and LINCOLN titled their 1998 thematic handbook Strategies of Qualitative Inquiry. In a second book by CRESWELL (1998), he uses the term "tradition" in place of analytical strategy. However, the point these authors make is identical. Specifically, the strategies—or traditions—advanced in these texts are much more than just "research methods." They consist of proven analytical strategies that have stood the test of time because they consistently produce high quality research. Each type of analytical strategy is linked to very specific assumptions about the nature of knowledge (ontology) and how such knowledge "can be known," accounted for or researched (epistemology). And these assumptions are, in turn, linked to vigorous debates about the most appropriate analytical strategies regarding specific types of topics, especially in the social world (CRESWELL 2002). However, following MILES and HUBERMAN (1994), CRESWELL (1998, 2002) and PUNCH (2000), I highlight the idea of contrasting analytical strategies because there are often several viable—and appropriate—ways to approach most research topics. Ultimately, when choosing amongst analytical strategies, it often boils down to a single question: "With respect to my research topic, does it (the analytical strategy) feel right?" [17]

A researcher's choices—if a choice exists—about analytical strategy often corresponds very closely with assumptions strongly associated with her or his beliefs about knowledge and the very best chance a researcher has at articulating something meaningful about their research topic. And thinking about this even further; at least with some researchers, will often reveal something about a specific research method—or set of methods—that makes good sense when considering a specific topic. "Lack of choice" refers to the situation when a researcher is given a specified research topic—and in some cases the analytical strategy they will use on that topic—by someone else. This happens in some instances, particularly in funded research and in disciplines or regions in which professors often specify topics for their students. [18]

In all types of research, having a good analytical strategy is important, but it is especially important in research topics concerning intercultural interaction. This is because it is often the case that the researcher is collecting data within settings where the culturally-based assumptions held by the researcher are very different than assumptions held by people in the research setting. A good analytical strategy is valuable in so far as many frameworks will illuminate these mismatching assumptions. [19]

Multiple methods is a term used in studies following a clear analytical strategy using two or more research methods. Examples of qualitative analytical strategies that commonly use multiple methods include the case study, grounded theory or ethnographic field studies (YIN, 2003; MILES & HUBERMAN, 1994; CRESWELL, 1998). [20]

While this protocol is highly complementary for use in mixed methods studies, that is, studies with an additional quantitative methodological component, this article will focus on the protocol's use in purely qualitative studies. [21]

In the next section, the interplay between analytically-based purposeful selection (CRESWELL, 2002) and theories of the middle range (MERTON, 1968) are highlighted. Robust analysis within many types of qualitative analytical strategies hinges on conceptually-driven identification and selection of research settings, data from these settings and especially the participants who will inform our research. [22]

This protocol draws its power to a great extent from strict attention to analytically based purposeful selection (CRESWELL, 2002). Purposeful selection means that research settings, data from these settings and research participants are located with reference to theory that indicates the selection is likely to address or lead to data that can answer research questions. Using theoretical concepts to guide the selection of data from the beginning—at the design stage of research—is the first step to interpretive data analysis designed to generate analytical generalizations—to theory. Analytical generalizations are very different than statistical generalizations—to populations—but are of no less value (YIN, 2002). Persons who do not understand the difference between analytical and statistical generalization often reveal themselves when asking how the results of a qualitative study can be used for direct generalizations to populations beyond the obvious scope of a research topic. Such questions miss the methodological point of analytical strategies in which statistical generalization is inappropriate, impossible, irrelevant—often all three. This is not the same as saying statistical generalizations are not important. They are—where statistical analysis has been used on a population and generalization to that population is claimed. The point here is that the goal of many types of analytical strategies using qualitative research methods is description, interpretation or explanation with regard to one or more research questions with reference to known theory. The most powerful answer to these questions, involves theory verification or theory generation (PUNCH, 2000). Theory verification involves interpreting or explaining a research question in terms of established theory, while theory generation involves interpretation or explanation of phenomena that has not been—up till the study—accounted for in terms of known theory (to the satisfaction of the researcher). Either way, qualitative researchers can often articulate original research findings—as long as there was an adequate purposeful selection done in the first place. But most importantly, an analytically based purposeful selection supplies the answer to the important questions: Why did you select those participants, that setting, this set of documents and not others? [23]

And to bring us full circle, robust analytical generalizations—to theory—are often an excellent basis for a statistical generalization—to a population. This is the precisely logic of many mixed methods designs. [24]

Theories of the middle range (MERTON, 1968) are theories and sets of concepts that do not claim or aim to explain a very wide scope of human activity, rather, they explain a limited scope extremely well. A practical example, drawn from my own research, involved the selection of interview participants in a study involving university faculty. Because the research topic had never been researched, specifically: The career potential of migrant scholars in Finnish universities (HOFFMAN, 2007); it was necessary to locate a group of scholars with a migrant background working in Finnish universities. I knew that locating the interview participants with regard to relevant theory would make robust analytical generalizations possible. Amongst the most useful theories I used was a body of work related to disciplinary cultures (BECHER & TROWLER, 2001). Particularly useful were their conceptual dimensions that mapped the cognitive territory on which all disciplines and their associated specialties can be located. Specifically: Hard and soft sciences versus pure and applied sciences. When used as intersecting continuums, four quadrants are formed, specifically hard-pure specialties, like theoretical mathematics and hard-applied specialties like mechanical engineering versus, soft-pure disciplines like sociology or soft-applied areas like education. By using this theoretical grid, I made sure that a conceptually based area was covered in participant selection and went on to discover important differences about the research topic that were in fact interpreted and later explained in terms of these same concepts. While these were not the only concepts used in the purposeful selection, they proved to be amongst the most important in the cited study because careful attention to theory insured conceptual variability amongst the participants. In plain terms, paying attention to this theory—along with other important theories—insured I didn't end up talking with too many similar types of faculty members. This example underlines the importance theoretical concepts play, both at the beginning of the research process, via purposeful selection, by illuminating participants that will inform a research topic. Just as important—in the interview data analysis stage—the use of the very same concepts may shine a bright light on the answer to a research question. The potential for illuminating new knowledge and/or intervention leverage points is much more likely when proceeding with a well-developed analytical strategy incorporating purposeful selection and theories of the middle range. [25]

The protocol introduced in this article highlights the communicative preferences participants have regarding types of interaction associated with different research methods. Before reviewing the types of research methods that work well together, the idea of communicative preferences is first presented. [26]

The idea of communicative preferences is rooted in practical advice given to my students in research methods courses. Specifically, regarding most efficient and effective way for masters-level and doctoral students to communicate with their research advisors in Finland, where some advisors keep extremely limited office hours. While teaching, student complaints about advisors—who never spent enough time advising their students—caused me to reflect on the direct observation that particular advisors had very clear communicative preferences which applied to almost everyone they dealt with, not just students. The practical advice I gave to my students was to systematically try at least three different approaches with their advisors and see which produced the best results: e-mail, via phone and face-to-face. I advised them to initially try e-mail with a very short request for a meeting or with specific questions, as this would allow their advisor to give a solid hint about his or her communicative preference. Once a clear communicative preference was established, the students were advised to continue to use the most effective communication channel they found. [27]

During the same period of time, I was conducting the earlier cited qualitative studies in which several different research methods were used, for example, thematic interviews, biographical e-mail correspondence and group interviews. During the course of these studies, I observed that research participants also had the same types of communicative preferences, although initially the difference wasn't as obvious. This was because each of the studies I was involved in normally employed one primary research method. However, because some of the studies dealt with similar topics, the advantages of approaching the same topic and research questions via fairly distinct methods—and incorporating the idea of communicative preferences—quickly became apparent. This was because of the type of interaction was so different in, for example a group interview, which was clearly optimal for some participants, while it clearly "shut down" meaningful participation for others. With one participant in particular—in a group interview—I had a hunch that s/he held very strong views about the topic under discussion, but for some reason did not voice these views during the interview. This hunch was shared by another member of the research team in the same interview. While it was very easy to speculate on several probable reasons for this, that's all we could do—speculate. We ended up discussing whether or not the participant in question might have responded very well to a follow-up e-mail query or individual interview. [28]

Another key element of this evolving approach was the quick proliferation of communication channels being adopted within society, especially those related to information and communication technology (ICT). It is quite clear that many people enjoy e-mail, blogs and instant messaging as much or more than face-to-face interaction or talking on the phone. And it is not difficult to locate peers or students who seem like "different people" depending on which medium—or language for that matter—they are encountered. Interviewing and other types of interactive data gathering via ICT-based interaction are also reflected in the growing literature on research methods (COUPER & HANSEN, 2002; ODIH, 2004; KAZMER & XIE, 2008). [29]

When taken together, it was clear that the confluence of these circumstances laid the foundation for the practical approach advanced here. [30]

While there are important advantages and disadvantages (for the researcher) associated with the use of different types of research methods within the same study, I will point out that if the object of the game is to stimulate participation in the research; using a variety of methods that corresponds to the communicative preferences of our potential participants, may be a better starting point than our own communicative preferences. This is because our own communicative preferences—as researchers—are often unarticulated—therefore not discussed. In other words a "choice" that research participants are stuck with: If they decide to participate in the study. This last point is emphasized because participants may often keen to participate in a study on a topic that is important to them, but may not be keen on the method a researcher is using to approach the topic: more precisely, the nature of interaction that a particular method entails. The point here: The choice of methods actually guarantees non-participation of some participants. This is the key rationale of approaching a topic with the idea of giving a participant a choice as to their mode of response. While the logic of this protocol will not address all research topics, it increases participation regarding many topics, a strong advantage over survey-based approaches. While surveys can be a powerful research method, suited to many research topics, many make the mistake of using them on topics that are not suitable, analytically speaking. When this is the case, research time is better spent on pursuing analytical generalizations that—in and of themselves—may be much more meaningful than half-baked statistical generalizations. Alternatively, solid analytical generalizations can provide a basis for statistical generalizations, if this is desirable. It depends on the topic and the goals of the research (HOFFMAN, 2005). [31]

Offering a participant several ways to contribute to the study—if possible—is a flexible approach to participation. This is especially true when dealing with "hard to access" participants, participants who are "very busy" and especially participants with good reasons for non-participation which they may be very reluctant to share (ADLER & ADLER, 2002; SHUY, 2004). [32]

The main idea of using the incremental interview approach protocol is based on two assertions. Firstly, different research participants have different potential with regard to the extent they will be able to contribute to the understanding of a research topic. Secondly, independent of what this potential is, participants may have clear communicative preferences. Both of these can and do influence the likelihood of participation in a study and the subsequent nature of that participation. By taking advantage of the fact we know these dynamics in advance, the protocol is used incrementally, or gradually, depending upon the process of locating participants, which often can yield significant data, as the example protocol indicates. But, in special focus here is the nature of the encounter the researcher has with each new participant, the extent to which that participant will inform the research and their communicative preferences. [33]

Regarding the nature of interaction, the reason the word incremental is stressed in this protocol is based on the author's recognition that—within the same study, it is possible to meet three participants whose interviews—for example, thematic interviews—each take 1.5 hours, but with very different outcomes. [34]

With Participant A, hardly anything is discussed that actually has to do with the research topic. There can be many reasons for this, ranging from poor purposeful selection on the part of the researcher(s), to the fact that some participants take a great deal of time in their verbal interactions (as do some researchers.) [35]

However, even when researchers have made poor choices and find themselves in interviews with participants they realize are not going to inform their research a great deal, no interview should be viewed as a waste of time. This is for several reasons. Firstly, if nothing else, these interviews force researchers to pay better attention to subsequent invitations to participants, reflecting on the reasons why they ended up speaking with someone—whose interview did not inform the research topic in general or research questions in particular. These types of interviews may turn out to be a blessing in disguise in so far as they force researchers to ask themselves if their analytical strategy is sound, specifically, if their topic, research questions and theoretical assumptions are congruent. Secondly, like all interviews, a particular participant—when listened to carefully—may illuminate issues that are actually more interesting than what the researchers thought it was they were investigating. Thirdly, practice. Specifically, there is no way to get good at interviewing—except by planning, facilitating and analyzing lots of interviews. Good interviewers never stop developing their skills. [36]

More important than the poor choices with regard to participant selection are the good ones— and the great ones. In the same study in which you encounter Participant A, who was somewhat of a blind alley, Participant B might turn out to be a person whose participation is essential for your research, a person who enables you to "connect the dots" of something very important. This connection may not have occurred except for your interview with that specific participant. [37]

And then there is Participant C. In some cases, the researcher might realize a particular participant is caught up in an absolutely unique set of circumstances with respect to the research topic (JOHNSON, 2002). It is easy to think—during these types of interviews—that a thematic interview won't even scratch the surface of what this participant has to say. It is during these interviews that researchers with a good awareness of their analytical options might be tempted to wonder if the individual they have encountered warrants re-thinking their approach to the research topic in some way, perhaps even questioning their most basic assumptions about the topic or the best way to approach it. [38]

Regarding communicative preferences, "similar types" of participants may have very different communicative preferences. One participant might be keen for a face-to-face interview. With a second participant, a phone call might be the only possibility you have for contact. A third might be much more comfortable first thinking and reflecting on the topic, then writing you a very detailed journal or diary entry that looks more like an essay or letter. This last type of data might take the form of an e-mail, blog entry or even be written long-hand. But all these forms of participation may answer your research questions and all allow response, follow-up and further interaction with the participants. A very practical point here about allowing for written responses: They eliminate the need for transcription—while allowing for interaction! [39]

While it may be easy to accept the idea of incorporating communicative preferences into your research, offering multiple participation options places extra demands on the researcher. These demands result from the fact that the methods of analysis associated with different kinds of data need to be thought-through, in advance, in order to make sure that all data a protocol links can be accounted for at all stages of analysis. [40]

It is for this reason that the careful selection of the analytical strategy, theories of the middle range and purposeful selection strategy needs to be coherent at the design stage of the research. Coherence will insure that your data—in the end—makes sense with respect to the topic specified, the research questions and the way in which research settings, data and especially research participants have been located. [41]

A second important practical point here: Most computer assisted qualitative data analysis software (CAQDAS) programs are set up for multiple data types and even assume the type of approach used in this data collection protocol. [42]

It is because of the variety of proven research methods we have at our disposal, the actual communicative preferences of research participants and the uncertainty even the most experienced researcher(s) will have with regard to the quality of their analytically-based purposeful selection, that the incorporation of a specific protocol is now advanced. The protocol explicitly links these known features of qualitative research to the unknown research settings we will encounter in our research and the nature of the interactions we will have with research participants within these settings. [43]

This protocol is a practical data collection framework. It links types of research methods that work well together at very specific stages of qualitative research. It assumes that the researcher already is following an establishing analytical strategy in which interviews, observations and documents may be important, like the case study, grounded theory or is conducting an ethnographic field study (MILES & HUBERMAN, 1994; YIN, 2003; CRESWELL, 1998). While the protocol highlights research methods that mutually reinforce each other, it offers no specific guidance completing the analysis of data—other than as a graphic illustration of the fact that many forms of qualitative analysis begin during the design stage of research. Neither can the protocol be used to predict which data will turn out to be the most important. This is because, firstly, several excellent textbooks, journal articles and courses already exist on data analysis. The utility of this protocol is in the explicit linkage(s) of the research methods within the framework. Secondly, the framework underlines a central feature that unites many qualitative analytical strategies: At the outset of research, the type of data that best answers a research question—especially a very interesting and original research question—may not become clear until well into the study. This is why the use of a tested analytical strategy, guided by solid linkages between theories of the middle range and purposeful selection with regard to one's focus on data, sets the researcher up for obtaining data that will answer their research questions. Keeping in mind what the protocol's purpose and limitations are, the following concrete example is advanced for consideration. [44]

The following example protocol is taken from an actual case study in progress by the author and his colleagues (HOFFMAN, KORHONEN & RAUNIO, 2008). The protocol relates stages of research, which draw attention to what is actually happening during that stage, but more importantly different research methods that are particularly suitable during each stage. Finally, the protocol explicitly draws attention to key decision points between stages at which the researcher(s) can ask themselves what the next steps are with regard to both data collection and participation. A particular feature of the protocol is that it draws attention to early opportunities that exist for data collection—which are often missed. It is these earliest opportunities, particularly at stages 1 and 2 in which very compelling data related to concepts guiding purposeful selection can be obtained. It should be emphasized that not all possible research methods are listed, only those most likely or relevant to a particular stage. In this sense, each protocol is different and depends on researcher preferences, skills, resources, settings and the research questions guiding the investigation amongst other things.

Table 1: Example incremental interview approach protocol (HOFFMAN, KORHONEN & RAUNIO, 2008) [45]

The use of different methods within established analytical strategies is nothing new. Researchers doing case studies, grounded theory and ethnographic field studies, among others, have long practiced the triangulation of data sources in order to produce credible results within these well-known traditions (CRESWELL, 1998). More recently, robust mixed-method strategies have been advanced for increasingly sophisticated designs that systematically allow for a great variety of creative approaches to research (CRESWELL, 2002; TASHAKKORI & TEDDLIE, 2003). [46]

Regarding the six stages of the example protocol above, the two most important caveats are: There are no absolute guidelines for approaching a research topic and there is always room for improvement. However, the most important feature of a protocol like this is that several features of the relationship between purposeful selection, data collection and research methods—within well-established analytical strategies—are rendered explicit with regard to the process of qualitative research. This is of particular importance with regard to the initial move from the pre-empirical stage of designing research to the empirical stage of doing research (PUNCH, 2000). [47]

This level of explicitness aims at analysis, from the very beginning of the research process. The protocol is an alternative to heading into the field, "doing a bunch of interviews," only to return with KVALE's 1000-page question: "How shall I find a method to analyze the 1,000 pages of interview transcripts I have collected?" (KVALE 1996, p.176) [48]

In addition, the protocol highlights several early opportunities for data collection that many interviewers miss. This is because the early stages, especially Stage 1 and Stage 2 have a great deal to do with articulating and operationalizing analytically based purposeful selection. This in turn, may confirm the researcher is on the right track. Alternatively, paying attention to data collection and associated reflection can set off alarm bells that something is not going quite as expected. In the analysis stage, many who use only interviews, without keeping good field notes or collecting other types of obvious data during the early empirical stages may have trouble articulating a convincing analysis during write-up. This is because the only data they have are their interviews, which may not—in and of themselves—answer their research questions completely. [49]

By focusing on the stages of research, the complementarities and relationships between different research methods are highlighted, while the decision points focus attention to the fact that there are continually options open for different types of participation—and non-participation—for both potential participants and the researcher. It is very practical to design "escape hatches" from the interview process because a wide variety of well known problems await anyone who does more than a few interviews (JOHNSON, 2002; ADLER & ADLER, 2002). [50]

The explicitness of reflection between stages—even if it is only for a few moments—draws attention to the time it takes us to ask ourselves: "Is what I thought was going to happen—at this stage of the research—actually happening?" This reflection happens on our own most of the time, but can also be discussed when working in teams. [51]

And when a participant "you don't want to lose" is waffling on participation, the idea of communicative preferences emphasizes the idea that participation—a challenge in many types of studies—can be enhanced by utilizing a broader spectrum of communication channels than many researchers have considered in the past. [52]

Finally, the protocol's rationale is based on the idea that approaching research topics a bit more explicitly will systematically turn up other interesting topics, that this is not unusual and even can be systematically taken into account. [53]

What remains to be done, in some cases, is developing sound analytical linkages from qualitative to quantitative research in mixed-methods designs. However, in many cases, robust qualitative analysis will address audience needs completely. It is hoped this protocol helps make this clear by underlining the complementary data sources qualitative researchers have at their disposal, in some cases. [54]

What this protocol will not do is relieve the researcher from the responsibility of developing the skills that only come from reading authoritative sources on analytical strategies and data analysis methods, seeking out instruction or coaching with regard to their use and developing the skill that only comes with experience: And applying these all to your research topic(s). One of the most practical uses of the protocol, however, is the realization that producing high quality qualitative studies is neither mythical nor particularly mysterious. It is mainly a question of continually developing and refining approaches to research, following up hunches and testing assumptions. [55]

Adler, Patricia A. & Adler, Peter (1998). Observational techniques. In Norman K. Denzin & Yvonna S. Lincoln (Eds.), Collecting and interpreting qualitative materials (pp.79-109). Thousand Oaks: Sage.

Adler, Patricia A. & Adler, Peter (2002). The reluctant respondent. In Jaber F. Gubrium & James A. Holstein (Eds.), Handbook of interview research: Context & method (pp.515-536). Thousand Oaks: Sage

Archer, Margaret (1995). Realist social theory: The morphogenetic approach. Cambridge: Cambridge University Press.

Becher, Tony & Trowler, Paul (2001). Academic tribes and territories: Intellectual enquiry and the culture of disciplines (2nd edition). Buckingham: Open University Press.

Brennan, John (2002). Transformation or reproduction? In Jürgen Enders & Oliver Fulton (Eds.), Higher education in a globalising world (pp.73-86). Dordrecht: Kluwer Academic Publishers.

Couper, Mick P. & Hansen, Sue Ellen (2002). Computer assisted interviewing. In Jaber F. Gubrium & James A. Holstein (Eds.), Handbook of interview research: Context & method (pp.557-575). Thousand Oaks: Sage.

Creswell, John (1998). Qualitative inquiry and research design: Choosing among five traditions. Thousand Oaks: Sage.

Creswell, John (2002). Research design: Qualitative, quantitative, and mixed method approaches. Thousand Oaks: Sage.

Denzin, Norman K. & Lincoln, Yvonna S. (Eds.) (1998). Strategies of qualitative inquiry. Thousand Oaks: Sage.

Hart, Chris (1998). Doing a literature review: Releasing the social science research imagination. London: Sage.

Hoffman, David M. (2005). The mobility of scientists and "best practices", or the wrong answers to the wrong questions? In Toni Gabaldón, Hugo Horta, Dagmar Meyer & José B. Pereira-Leal (Eds.), Career paths and mobility of researchers in Europe (pp.86-97). Göttingen: Cuvillier Verlag.

Hoffman, David M. (2007). The career potential of migrant scholars: A multiple case study of long-term academic mobility in Finnish universities. Higher Education in Europe, 11(4), 317-331.

Hoffman, David M. (in press). Changing academic mobility patterns and international migration: What will academic mobility mean in the 21st century? Journal of Studies on International Education.

Hoffman, David M.; Korhonen, Marjaana & Raunio, Mika (2008). Is internationalization worth it? Finnish universities as international academic environments. Presentation of case study protocol at 10th Anniversary Symposium on Higher Education, Institute of Educational Research, University of Jyväskylä, Finland.

Hoffman, David M.; Välimaa, Jussi & Huusko, Mira (2008). The Bologna process in academic basic units: Finnish universities and competitive horizons. In Jussi Välimaa & Oili-Helena Ylijoki (Eds.), Cultural perspectives on higher education (pp.221-237). Berlin: Springer.

Johnson, John M. (2002). In-depth interviewing. In Jaber F. Gubrium & James A. Holstein (Eds.), Handbook of interview research: Context & method (pp.103-119). Thousand Oaks: Sage.

Kazmer, Michelle & Xie, Bo (2008). Qualitative interviewing in internet studies: Playing with the method. Information, Communication & Society, 11(2), 257-278.

Kuhn, Thomas S. (1996). The structure of scientific revolutions (3rd revised edition). Chicago: Chicago University Press.

Kvale, Steinar (1996). Interviews: An introduction to qualitative research interviewing. London: Sage.

Marginson, Simon & van der Wende, Marijk (2007). Globalisation and higher education. Education Working Paper, 8. Paris: OECD

Merton, Robert (1968). Social theory and social structure (enlarged edition). New York: Free Press.

Miles, Matthew B. & Huberman, A. Michael (1994). Qualitative data analysis: An expanded sourcebook (2nd edition). Thousand Oaks: Sage.

Odih, Pamelah (2004). Using the Internet. In Clive Seale (Ed.), Researching culture and society (pp.279-291). London: Sage.

Seale, Clive (2004). My research practice. In Clive Seale (Ed.), Researching culture and society (pp.464-473). London: Sage.

Shuy, Roger (2004). In-person versus telephone interviewing. In Jaber F. Gubrium & James A. Holstein (Eds.), Handbook of interview research: Context & method (pp.537-555). Thousand Oaks: Sage.

Tashakkori, Abbas & Teddlie, Charles (Eds.) (2003). Handbook of mixed methods in social & behavioral research. Thousand Oaks: Sage.

Yin, Robert K. (2003). Case study research: Design and methods (3rd edition). Newbury Park: Sage.

Dr. David M. HOFFMAN is a researcher at the Institution of Educational Research at the University of Jyväskylä, Finland. His research focuses on international comparative higher education, international migration and ethnic relations and intercultural communication. He regularly teaches courses in intercultural communication, higher education and research methods in Finland, France and Canada.

Higher Education Studies Research Team
P.O. Box 35 (KTL)
40014 University of Jyväskylä
Finland

Hoffman, David M. (2009). Multiple Methods, Communicative Preferences and the Incremental Interview Approach Protocol [55 paragraphs]. Forum Qualitative Sozialforschung / Forum: Qualitative Social Research, 10(1), Art. 41, http://nbn-resolving.de/urn:nbn:de:0114-fqs0901419.

Forum Qualitative Sozialforschung / Forum: Qualitative Social Research (FQS)

ISSN 1438-5627

Creative Commons Attribution 4.0 International License

Incremental Stages	What is Happening	Primary Research Method Options at this Stage
Stage 1: Transition from design stage of research to purposeful selection of settings, data and research participants.	The research plan written, with a clear analytical strategy, purposeful selection of settings, data and participants begins	* Direct observation * Participant observation * Documentation
Decision point (yes/no)	Researcher(s) assess(es) initial data: Is design adequate? - If "yes," then Stage 2, - If "no," redesign	* Research log, journal or notes regarding progress. * Meeting with research team to assess progress
Stage 2: 100% of potential research participants contacted	Initial contact, based on analytical purposeful selection with short request for participation.	* Direct observation * Participant observation * Documentation
Decision point (yes/no)	*Does further participation seem like a good idea or possible? - For participants who agree, proceed to Stage 3.	* Research log, journal or notes regarding progress. * Meeting with research team to assess progress
Stage 3: 100% of participants who initially agree to participate) * Gather full contact information for participant, * Send information about research, consent and participation options	* Participation begins. Participant is fully informed about the nature of the research, nature of consent and options for participation * Biographical information is collected	* Direct or participant observation * Direct observation * Participant observation * Documentation * Biographical data collection
Decision point (yes/no)	*Is further participation warranted? - For clear "Type 1 participants" participation is terminated here. - For "Types 2 & 3," continue	* Research log, journal or notes regarding progress. * Meeting with research team to assess progress
Stage 4 (Unknown % of participants originally opting for participation) Participant/researcher interaction for purposes of gathering data	* Interview takes place and/or journal is e-mailed, mailed, then collected * Interview concluded, journal collected	* Interview * Biographically-based methods * Direct or participant observation * Documentation
Decision point (yes/no)	* Is further participation warranted? - For "Type 2 participants" participation is terminated here with thanks & request for participation in follow-up studies. - If "Type 3 participant," further participation is requested.	* Research log, journal or notes regarding progress. * Meeting with research team to assess progress
Stage 5 (For Unknown % of participants)	Unique case is located, additional data collection warranted. Following data collection, participation is terminated with thanks & request for participation in follow-up study.	* In-depth interview if possible * Journal provided * Supplementary documentation requested
Decision point (yes/no)	* Has enough data been collected to answer research questions? - If yes, then analyze data and write-up study - If no, continue data collection.	* Research log, journal or notes regarding progress. * Meeting with research team to assess progress
Stage 6	Member check, Write-up	* Writing process
Decision point (yes/no)	* Has a new study been identified? If yes, plan study.	* Planning stage