1999 AERC Proceedings

Constructing "Composite Dialogs" from Qualitative Data: Towards Representing and Managing Diverse Perspectives

Peter G. Malvicini
Cornell University

Abstract: During empirical research on adult learning, I developed a technical method of constructing dialogs from many interviews as an approach to qualitative data. Composite Dialogs present diverse perspectives while preserving elements of context and affect often lost through other methods. The approach should be of particular interest to adult educators who desire to link research to social action for democratic change.

Fieldwork, Data, and Misperceptions. I returned from nearly a year of intensive fieldwork in the Philippines with complete transcripts of 53 individual and focus group interviews and a file drawer full of documents. I had gone to the Philippines to facilitate action research, studying the learning processes involved when people significantly change their approaches to development work. I examined processes of transformation from centralized "blueprint" approaches, to more participatory, community-based development efforts in the context of extensive government decentralization. I returned confident that the efforts had been of some benefit to my partners. However, the daunting task of analyzing, interpreting and presenting over five hundred pages of data in a way that was meaningful and useful, both to "the field" and to practitioners in the field, nearly paralyzed me. In prior research, I had built category systems for qualitative data, but this approach did not seem appropriate now. How could I reduce my material to a series of disembodied sound bites? Not this time. I was groping for a way to respect the richness of the source material, yet present findings in a manageable form.

I recognized that, to some extent, my research questions were misperceptions I brought to the field. However, as I continued to immerse myself in the data both in the field and at home, new understandings emerged, replacing old hypotheses with better questions. My data did not answer these questions conclusively but provided a range of answers and discussion around these questions. This paper outlines an approach I developed to address my concerns. The approach is concurrently analytic, interpretive and representational.

A Theory-Method Gap in Adult Education Research. Adult education research benefits from a variety of qualitative methods borrowed from other fields (Merriam, 1989, 1998). However, prevalent qualitative research methods sometimes reduce the complexity of experiences, ignore affective dimensions altogether, or decontextualize results in a way that sacrifices meaning. Taylor's critical review of research on transformative learning (1997) elucidates some of these weaknesses calling for different approaches to research. In the traditions of critical theory, action research, and feminist epistemology, some methods clash with the values of popular participation or fair representation of multiple perspectives (Hall 1996; Joyappa & Martin, 1996; Cocklin, 1996). While examining adult learning, I developed a technical method of constructing dialogs as an approach to qualitative data. This was not driven by a desire to use a novel form of representation. I constructed dialogs because I was dissatisfied with other methods of qualitative analysis. This approach generates knowledge by revealing and highlighting the differences in people=s perspectives in a way that begs action.

In my own research on adult learning, dialogs help (1) manage the diversity of participant's experiences; (2) reveal the complex nature of adult learning; and (3) suggest the possibility of organizing to support learning and change. From the dialogs, I drew out four different arenas for taking action, including opportunities for: (1) creating spaces for learning and change; (2) examining assumptions about the local capacity for change; (3) learning through participatory action; and (4) democratizing bureaucratic organizations.

Perspective. Throughout my research, tension remains between building theory and informing practice. The framework of Participatory Action Research helps me to preserve that tension (Greenwood & Levin, 1998; Fals Borda & Rahman, 1991). Guba & Lincoln's (1989) trustworthiness criteria provide a vocabulary to discuss the quality and credibility of the dialogs. The dialogs should catalyze reflection and guide readers toward action. To the extent they are effective in doing so, they demonstrate their authenticity.

Observations on Constructing Dialogs. The dialogs began as an analytic approach to deal with a vast amount of interview data. Much qualitative work could be better understood at this level of analysis, perhaps more so than through descriptive categories or third-person ethnographies. The dialogs keep the reader close to the data while preserving some of the complexity and affect of the original material. Constructed dialogs, though not without limitations, have some distinct advantages. They can: (1) present "silenced" voices, (2) bring perspectives from participants who are unable to come together either because of physical or ideological distance, (3) and avoid the distortions of negative group dynamics in a live discussion.

Researchers are quick to ask practical questions, such as: How long did this take? Figure 1 presents a funnel illustrating the time it took me to both develop the approach and apply it to my data. Over time, I became more facile with the technique. The first dialog took me over a month to construct, with many false starts, while the fourth dialog took just over a week. Furthermore, I constructed dialogs retrospectively. If one planned to use the approach, it would likely move more quickly. Figure 1 also shows how the dialogs economize on the original source material. In my case, I used roughly 12% of the original material in the dialogs.

Sample Section of Dialog. Box 1 is a section of dialog (approximately one tenth of the entire dialog) that explores the process of change in the way people view and approach development work and their role in it. I tried to understand people's concepts and ideas about development work and what difference they make to their own practice. I also tried to understand the process, how these changes came about. This particular dialog sheds light on the process of transformative learning.

In this section of dialog (Box 1), Pia, Jaime and Joel share the "how" of their changes in thinking with personal stories. They talk about experiences rooted in different contexts (family, hometown, school, job, etc.) over a period of almost thirty years. They move from simply sharing experiences to reflecting about them aloud. Profound changes did occur. Participants in the complete dialog demonstrate varying degrees of critical reflection, examining and questioning their assumptions and past learning about development. They share direct experience in taking action, particularly when that action confronted power structures that were, in their opinion, barriers to development or democracy. Similarly, direct experience working with the poor has changed the way people view themselves. Many illustrations are personal and often emotionally charged. Transformation is also put forth as a social process worked out through sharing these powerful experiences and processing them through discussions with others, be they colleagues or teachers or family members. The dialog presents a range of different views, from development as empowerment to development as the creation of infrastructure or the delivery of services.

Did the dialogs actually occur? No (and yes). They are a construction or representation of many different interviews, thus the term "composite." The excerpts in the dialogs are the actual words, verbatim; I have edited the dialogs only for length and clarity. It may be useful to envision the dialogs in terms of a "reader's theater" where actors speak into the center of the room rather than having a conversation that goes back and forth directly between individual participants.

Strength of Approach. Constructing dialogs provided new insight into the learning processes that I was studying. The different participants share many ideas and illustrate them with their personal experiences. The dialogs provide a new way or platform for these perspectives to interact. This made the analytic process more complex, dimensional, and political. The dialog approach makes it difficult to avoid issues of power and authority as they appear in the data.

Theories, while usually good at describing the present situation or the ideal, often leave the practical problem of how to get there or "what do we do next?" unanswered (Toulmin, 1996). In contrast to theories, the dialogs express concrete alternatives for taking action. The 'how' problem becomes not only discussible, but the participants recognize what is standing in the way of change and start to remove the barriers.

Implications. The dialogs seek to preserve the diversity of the conversation (Flood & Romm, 1997). Each perspective carries only a piece of the conclusion. Some analytic approaches reduce material to a diluted consensus or majority voice, presenting an oversimplified picture that gives researchers a false sense of security in a narrow set of conclusions--complexity is much more unsettling, but a far more accurate and useful representation. Dialogs present no forced consensus; they are comfortable in doing so. By creating the space to share perspectives and approaches, they create an informed consensus of action rather than a false consensus of ideas. This does not dismiss the need for the discussion of ideas -- far from it. It allows participants the freedom to discuss their ideas while lessening the fears of a paralyzing ideological stalemate that would stifle any action.

Composite Dialogs Step-By-Step. In Figure 2 below, I outline the steps I used to construct dialogs from semi-structured interviews. (1) Interview data were imported from word processing files into FolioViews® software as a textual database. (2) I coded the data descriptively based on my original research questions. (3) From this coding, several topics emerged as broad themes. I then turned these topics into dialog questions. For example, I turned the topic "political economy of transformation" into the question, "What is preventing you from putting your view into practice?" (4) Using the software's search engine, I identified excerpts and coded them according to the appropriate dialog question. These excerpts became the raw material for each dialog. (5) At this point, the excerpts for a particular dialog question were re-read and sorted for similarity. The results formed the subtopics for a given dialog, i.e. "political constraints," "lack of shared understanding," etc. This took several iterations. (6) I then labeled each subtopic (group of excerpts) and used a flow-diagram program to map out the rhythm and sequence of the particular dialog. (7) From the organized excerpts, I looked for the different ways people approached the question, and then constructed three different dialogic positions. (8) I laid out the dialog by further sorting the excerpts into these different positions. (9) At this point, I removed some excerpts because they were redundant, no longer fit the dialog, or were less significant. I tried to treat each position fairly, essentially applying "ideal conditions for discourse" to the dialogs I was constructing by avoiding a dominant position. Constructing dialogs was both a demanding analytic process and an intensely creative one. Arranging and rearranging text was necessary to achieve a dialog that allowed the reader to follow the flow of the discussion. With each subsequent dialog, the method became more natural.

The diagram in Figure 2 is deceptive in that it makes the process appear linear. In reality, there was significant iteration through hypotheses, methods, conclusions, and action. The interplay between these elements was intense as I tacked back and forth between each, generating new conclusions that superceded prior conceptions. As one goes down the diagram, the analysis gets more sophisticated yet remains grounded in the empirical data.

The existence of a piecemeal world of works-in-progress is messy and difficult. Dialogs provide a strategy to manage it. In my research, action implications did not emerge from a single perspective but flow from the dialog among the perspectives. The dialogs preserve the complexity of the source material while presenting the diversity of results in an accessible form. Alternative perspectives are not dismissed or prevented from entering the discourse.

Future research efforts in adult education could benefit from constructing, interpreting, and acting on dialogs collaboratively with participants. With adaptation, the dialogs lend themselves to performance in the spirit of popular theater (Bates, 1996), or could serve as a basis for quantitative surveys. Composite Dialogs are but one strategy to construct narratives that are more authentic.

Bates, R. A. (1996). Popular Theater: A Useful Process for Adult Educators. Adult Education Quarterly, 46(4), 224-236.

[Home] [Search] [Conference Proceedings] [Back] [Top]