One of the most important decisions that I made about methods in the early stages of my research was that I wanted to conduct interview-based research. My academic training as a sociologist and my experience as a researcher have given me previous opportunities to experiment with effective interview technique, as well as to develop an appreciation for the strengths of this method of data collection. While qualitative research methods have a long history within the social sciences, they have always lacked the status of more quantitative research approaches (Brockington & Sullivan, 2003:59; Flick, 2006:35; Mason, 2002:1). However, the challenges to positivism posed by feminist, anti-racist, and poststructuralist critiques of modernist social science have led to increased in interest in the potential of these methods to enable us to be more flexible, creative, and inclusive in our approach to data collection (Denzin & Lincoln, 2003b). While there is considerable diversity within qualitative research approaches, with some more wedded to the positivist paradigm than others (Denzin & Lincoln, 2003b:14-15), the common commitment to ‘inductive’ reasoning, and the focus on capturing the
specificity of individual experience have provided useful opportunities for those researchers interested in resisting easy generalisation in favour of the construction of richer, more complex understandings of social life (Mason, 2002:1; Tolich & Davidson, 1999:34).
Methodologically, the strength of qualitative research lies in the opportunities it provides to produce more in-depth understandings of the complexity of people’s lives. Qualitative research starts with the assumption that people are valuable sources of information about their own experiences, and that “much can be learned from direct, extended conversations with individuals whose thoughts and opinions are critical for understanding a topic” (Vaughn et al., 1996:17). Qualitative researchers begin with the daily existence of individuals, focusing on the ways that explorations of personal experience can enable us to build more comprehensive pictures of social reality. As Cole and Knowles (2001:11) sum up:
[C]lusters of individual lives make up communities, societies, and cultures. To understand some of the complexities, complications and confusions within the life of just one member of a community is to gain insights into the collective.
As well as having high methodological ‘validity’, qualitative methods also have distinct political and epistemological advantages. The emphasis on flexibility and the need to remain open to the “ambience of the exchange” (Rowles & Reinharz, 1988:6), provide increased opportunities for participants to have input into the direction of research. This flexibility can also promote greater transparency around the ways that data is actively ‘constructed’ by researcher and participant at a given time and place under specific material conditions. In this way, qualitative methods can make it easier to ‘situate’ our research efforts, as well as to “preserve the presence of the active and experiencing subject” (Smith, 1987:105).
Finally, the dialectical relationship between data collection and analysis within qualitative research is also useful for challenging the traditional over-determinism of theory within quantitative hypothesis testing approaches. Qualitative researchers argue against making a “parade of theory” (Firth in Wolcott, 2001:76); asserting, instead, that data collection and analysis should be intertwined and develop together according to the unique conditions experienced within a research project (Berg, 1998; Smith, 2002:27). As Smith
(1987:11) contends, the decision to adopt a more naturalistic approach to data collection involves “renouncing theoretical projects that seek full development and coherence prior to an encounter with the world”. Such an approach is consistent with Gibson-Graham’s (2006) argument in favour of more tentative ‘weak’ development theorising. Qualitative researchers seek to avoid the arrogance of positivist overgeneralisation. As Wolcott (2001) cautions:
By the very nature of the way we approach things – flatfooted observers with feet of clay – we tend at most to be theory borrowers (or theory poachers, as others may see us) rather than theory builders. Taking a model of theory-driven research derived from the so-called hard sciences doesn’t serve anything but our already heightened sense of physics envy. Unless you think one must wear a white lab coat to be a careful observer, forget that model and keep your theorizing modest and relevant. (Wolcott, 2001:77)
Despite the usefulness of qualitative research approaches, however, there are distinct challenges associated with the decision to use them within a development studies research project. A number of researchers have documented the ways in which the policy focus of development research – its emphasis on identifying mechanisms for inducing social change – continues to limit the scope, creativity, and methodological veracity of qualitative research projects (Hausner, 2006:323; Robinson-Pant, 2001:166). These authors assert that the focus on policy applicability has created a bias in favour of quantitative methods, because these methods provide numerical values that can be easily translated into policy goals (Hausner, 2006:323; Robinson-Pant, 2001:166). Amongst many within the development industry, Brockington and Sullivan (2003:59) argue, qualitative methods still have “a reputation for being anecdotal or associated with ideas that cannot be described with hard, secure facts”.
While the use of qualitative methods is slowly increasing in popularity in the development industry, due to the opportunities these methods offer to produce research on more egalitarian terms with participants, the bias in favour of quantitative methods remains. This is most evident in the continuing demands upon qualitative development researchers to meet the conditions of research reliability and validity established within the positivist quantitative tradition. An important example of this is the ongoing perception that ‘reflexivity’ should be avoided in development research reports because it can be seen to represent a lack of confidence in the validity of data (Robinson-Pant, 2001:167). Development researcher and ethnographer Sondra Hausner (2006:231),
commenting on this phenomenon, argues that while anthropology, as a critical social science, “insists that some self-revelation is critical to its method”, development studies, unfortunately, still often “requires that self-critique be an internal or private matter only”. Literacy researcher Anne Robinson-Pant (2001:167) contends that the other important evidence of the positivist, quantitative bias in development research is the restriction qualitative researchers face on changing research design/questions as research evolves for fear that flexibility is ‘unscientific’.
However, qualitative researchers from within traditional social science disciplines such as anthropology and sociology have become increasingly vocal in their opposition to the ways in which open-ended qualitative research is devalued within development practice. They argue that while outcomes, indicators, and questionnaires can be important tools within evaluation research, they are also problematic because “they presuppose the category. You know what you are looking for before you get there” (Hausner, 2006:337). There are considerable advantages associated with using open-ended research approaches. They can provide a necessary degree of flexibility that is “fundamental to the changing nature of human relationships” (Hausner, 2006:323) and, therefore, to efforts to capture the diversity and dynamism of social life. They can also provide greater opportunities to be responsive to the priorities of participants.
As Robinson Pant (2001), reflecting on her own experience as an ethnographer involved in evaluating a woman’s literacy project, asserts, ultimately the choice to use qualitative methods is not only about collecting better data, it is about taking a more creative approach that ensures all possible solutions to problems, not just those currently on offer, are considered. It implies, she argues, “a decision to take a more holistic approach to programming and possibly completely chang[ing] [the] overall direction” of a project (Robinson-Pant, 2001:166). Fellow anthropologist Hausner (2006) reinforces the power of qualitative research – in this case, ethnography - to lead to better, more effective, and more enduring development programmes:
Ethnography can tell programmers stories they did not know existed, demonstrate links and connections that no questionnaire could have dreamed up, and establish personal connections that can make the difference between a development project that everyone wants to milk and one that everyone wants to participate in. … Ethnography allows you to open the field, and direct your first line of inquiry to those for whom you intend your project, not to the rule
book that governs program design. So you know what is working and what is not, in your informants’ own language. The potential of open-ended research – ethnography – that does not presuppose the categories of inquiry is enormous. (Hausner, 2006:338-9)
French anthropologist Olivier de Sardan (2005) also extols the value of traditional anthropological or qualitative sociological research to increase the effectiveness of development programming. He asserts that in order to develop effective solutions to development problems we need to take advantage of the wealth of experience these disciplines have in designing methodologically rigorous research that ensures that we get a holistic picture of the social reality we seek to ‘amend’. Olivier de Sardan (2005:26) argues that social research requires “a level of competence that cannot be improvised”, and is increasingly being compromised by the policy imperative within development research:
The confrontation of varied social logics surrounding development projects constitutes a complex social phenomenon which economists, agronomists and decisions makers tend to ignore. In the face of the recurrent gap between expected behaviour and real behaviour, in the face of the deviations to which all development operations are subject, in consequence of the reactions of target groups, developers tend to resort to pseudo-sociological notions that bear a closer resemblance to clichés and stereotypes than to analytical tools. … Lacksadaisacalreferences to ‘cultural factors’ are more often than not oblivious not only of the existence of sub-cultures and internal cultural diversity within a given social group, but also of the influence of social cleavages (age, sex, social classes, among others) and of norms and behaviour. They lose sight of the fact that ‘culture’ is a construct subjected to continuous syncretic processes, and the object of symbolic struggles. (Olivier de Sardan, 2005:26)
Olivier de Sardan (2005) maintains that the future of effective development intervention will be dependent on our ability to eschew this “do-it-yourself sociology” (Olivier de Sardan, 2005:28) in favour of a return to a more rigorous qualitative research programme that starts with lives rather than projects. The following section explores the parameters of such a programme. Such a project could, optimistically attempt to build on the strengths of the more traditional conventions of research ‘rigour’ within the qualitative paradigm, as well as exploring the usefulness of some of the more recent feminist, anti- racist and postmodernist attempts to build more critical, reflexive qualitative methods.