DEBATE DEBATE

Debate sobre o artigo de Minayo & Sanches

Debate on the paper by Minayo & Sanches

Dominique Behague

Centro de Pesquisas Epidemiológicas
Universidade Federal de Pelotas

This paper has touched upon a series of provocative topics which cannot be criticized as a whole here. After such a careful development of the historical roots of quantitative and qualitative theories and methods, it may now be more fruitful to focus on the very question that the paper originally set out to answer: Do these methods oppose or complement each other? What are the difficulties that arise when both methods are used concurrently in a single project and how can researchers, on a practical basis, work to make them complement each other?

The problem is not easily solved by simply stating that researchers should relativize their positions and realize that neither method is complete for the understanding of reality. Too often, vast amounts of qualitative and quantitative data lie side by side without ever being actively used to direct research questions and objectives. Before productive interdisciplinary collaboration can occur, researchers must first recognize the culture of disciplinary "professionalism" and theoretical specialization in which they have been indoctrinated. In part, it is this "culture" which renders both methods inaccessible to the uninitiated and excessively "purist", thereby inhibiting adaptability to the needs of particular research settings.

Although the authors state that they prefer to generalize the discussion, I believe the root of the difficulty of integrating quantitative and qualitative goals lies in the lack of explicitly stated and commonly held research objectives and goals. In the field of public health, for example, much data is collected which theoretically and hypothetically could be beneficial for future project implementations but which, in reality, is of little use. "Problem-oriented" research, on the other hand, facilitates the integration of qualitative and quantitative methods, precisely because it is framed to provide specific solutions in particular cultural settings rather than to contribute to generalized theoretical concerns (Boehm, 1982).

The authors state that qualitative methods serve to investigate in greater depth questions raised by quantitative studies, as well as to raise questions to be studied by statistical methods in subsequent stages. Specifically, a preliminary qualitative research phase can be useful in 1) structuring questionnaires by phrasing questions in culturally appropriate ways which are more likely to elicit answers which are more proximate to respondent's uninhibited interpretations; and in 2) discovering unexpected knowledge or patterns which may otherwise not be incorporated in a survey.

This sort of collaboration requires a commitment to continually searching for innovative methods which adapt to the needs of a particular research area or problem. However, it carries an additional challenge: Is the burden of applying data to the "other side" to be placed on the lay person? How do we make the data accessible to those who are not specialists? Some have called for a "deprofessionalization" of the social sciences to ensure both multidisciplinary research and more community involvement in the research process (Nichter, 1984).

The supremacy of quantitative methods over qualitative methods in the social scientific fields have lured some anthropologists into highly complex attempts at quantifying such variables as social power, status, and psychological intentions. Is this approach answering the questions outlined by the needs of particular programs? Furthermore, is this approach reaping the benefits of the strengths which qualitative methods have to offer?

While quantitative research can adequately define macrolevel social processes, qualitative methods are best suited to investigate microlevel patterns in local settings. Conceptually, we are lacking methods which help us link microlevel patterns with macrolevel outcomes. Qualitative studies, which must predefine variables and specify relationships a priori, do not allow for the discovery of new knowledge or for the investigation of deeper questions concerning how these relationships are played out. For example, several studies have shown a statistically significant association between maternal education and child health. However, we are left wondering how one year of schooling "actually influences breast-feeding, if at al" (Mosley, 1992).

These sorts of questions are important for the often implicit and forgotten research objective of many projects: to make policy recommendations and to aid in the implementation of health promotion programs. Policy-makers are rarely persuaded by studies which are not based on numbers. Yet the power of numbers is to show associations. Once we know that two variables are statistically associated, how do we proceed to understand how the relationship between these two variables actually came about?