Social Indicators Research

, 90:5

Wellbeing Research in Developing Countries: Reviewing the Role of Qualitative Methods

Authors

    • Young LivesDepartment of International Development
    • Wellbeing in Developing Countries ESRC Research GroupUniversity of Bath
  • Gina Crivello
    • Young Lives, University of Oxford
  • Martin Woodhead
    • Young Lives, University of Oxford
    • Child and Youth Studies Group, The Open University
Article

DOI: 10.1007/s11205-008-9310-z

Cite this article as:
Camfield, L., Crivello, G. & Woodhead, M. Soc Indic Res (2009) 90: 5. doi:10.1007/s11205-008-9310-z

Abstract

The authors review the contribution of qualitative methods to exploring concepts and experiences of wellbeing among children and adults living in developing countries. They provide examples illustrating the potential of these methods for gaining a holistic and contextual understanding of people’s perceptions and experiences. Some of these come from Young Lives, an innovative long-term international research project investigating the changing nature of child poverty in India, Ethiopia, Peru and Vietnam (http://www.younglives.org.uk), and others from the Wellbeing in Developing Countries ESRC research group (WeD), an international, inter-disciplinary project exploring the social and cultural construction of wellbeing in Bangladesh, Ethiopia, Peru and Thailand (http://www.welldev.org.uk). The authors show how qualitative methods can be used both alongside and as part of the development of sensitive and relevant quantitative measures, and provide some practical and methodological recommendations. They propose that qualitative approaches are essential in understanding people’s experiences of wellbeing, both now and in the future. However, the authors caution that while these offer many benefits, for example, a less structured and hierarchical engagement between researcher and participant; they require time, energy, and sensitivity. Qualitative methods also work best when used by trained and experienced researchers working in the local language/s in a community where some rapport has already been established. Finally, the paper recommends combining data from qualitative and quantitative approaches (e.g. psychological measures or household surveys) to enhance its explanatory power.

Keywords

WellbeingMixed methodsQualitativeDeveloping countriesMethodological

1 Introduction

Openness to insights from other disciplines1 has been the hallmark of organizations such as International Society for Quality of Life Research, reflecting the fluid and multidimensional nature of its object of study. The purpose of this paper is to illustrate through vivid example the value of qualitative approaches to researchers, policy makers, and practitioners who construct or use social indicators to map quality of life, with the ultimate goal of enhancing the wellbeing of people in developing countries. It provides a necessarily selective mapping of the terrain for future exploration and aims to inspire readers to engage with qualitative approaches to and literatures on wellbeing, even if they do not adopt qualitative methods. The approaches described in the paper have particular intellectual histories that we cannot do justice to here; however, this should not prevent a pragmatic recognition of their potential as a resource within an integrated investigation of quality of life or wellbeing. For example, the role of participatory methods in understanding, and possibly even developing indicators of wellbeing is illustrated with studies exploring understandings of poverty, illbeing, and vulnerability; wellbeing; and resilience.

Concepts of quality of life and wellbeing, which for the purposes of this paper are treated as synonymous enable engagement with the whole of people’s lives and provide more accurate representations and measures than approaches that focus explicitly or implicitly on a single dimension (for example, health or income). However, there are multiple definitions of these terms that reflect different philosophical traditions, and little consensus, even within disciplines. For example, wellbeing can be used to refer to any or all of the following, all of which have different implications for research or intervention: a subjective experience or state of being (Diener 1984); the space where wellbeing can or should occur (Sen 1990) or a process with wellbeing as its goal (Aristotle, 350 BC); and, after Veenhoven (2000), the ‘liveability’ of the environment and the ‘life ability’ of the person. While definitions of wellbeing are contested (it is tempting to succumb to the authoritative pessimism of Hird that ‘there is no accepted definition of wellbeing’ [2003, p. 4]), there are some common understandings. These were reflected in McAllister’s (2005) recent review of the wellbeing literature which defined wellbeing as “more than the absence of illness or pathology […with] subjective (self-assessed) and objective (ascribed) dimensions; it can be measured at the level of individuals or society; it accounts for elements of life satisfaction that cannot be defined, explained or primarily influenced by economic growth” (ibid, p. 2).

In the context of research in developing countries these subtleties become more apparent—how do local people understand wellbeing and how do these understandings vary according to life phase, gender, socio-economic status, etc., within a single community? How do people pursue what they see as wellbeing or a good life and what trade offs are they required to make to attain or preserve it? (e.g. between self and household, present and future) What resources can people draw on in their pursuit of a good life for themselves, their families, and communities, and what are the political and social barriers? There is a further question of how far development agencies should take people’s priorities, values and visions of wellbeing into account to ensure the credibility and legitimacy of development (Copestake 2007). Additionally, there is growing interest in how public conceptualizations of wellbeing are constructed (Deneulin and Townsend 2007), especially in relation to children (Hood 2007).

These concerns are part of a shift within international development and child indicators research (e.g. Camfield and McGregor 2005; Ben-Arieh 2006) from a deficit view that focuses on survival, to one that acknowledges people’s resources and agency and pursuit of wellbeing. The contribution of qualitative approaches to a focus on people’s resources and agency is that they can encompass areas of people’s lives that are influential and important, but rarely measured (for example, spirituality and religious practice). More participatory approaches also challenge the reliance on experts and proxies by taking a subjective approach to people’s experiences and treating them as researchers as well as subjects. We propose that exploring understandings and experiences of wellbeing using qualitative research is valuable in its own right, as the paper illustrates, and also improves the accuracy of measurement. Qualitative research can make measures more comprehensible and relevant to respondents, provide contextual information to explain particular outcomes, and most importantly, ensure that the ‘stylised facts’ such as the ‘a dollar a day’ metric that influence international assistance are based on measures of what matters.

The paper begins with a brief introduction to the important but slippery concept of wellbeing, describing types of ‘wellbeing research’ that have been undertaken with adults and children. This is necessarily selective as wellbeing research is a broad category: most international poverty research claims to be about wellbeing (e.g. Coudouel et al. 2001), and studies of children’s development are implicitly about their wellbeing. The paper then introduces qualitative approaches—what they are and how they can be helpful—before focusing on two overlapping approaches to research: qualitative (for example, ethnographic, participatory) and mixed method, which combines insights from qualitative and quantitative. We have ordered the paper in this way as a heuristic device; boundaries between the approaches are blurred on paper and in practice, and there is no ‘grand narrative’. We have also tried to resist the temptation to compare ideal types rather than messy realities (for example, the lone anthropologist working for 30 years in a single community versus a team of participatory action researchers interviewing villagers over the course of a week). A reality check is provided by including some examples from our own research with the Wellbeing in Developing Countries ESRC research group (WeD), an international, inter-disciplinary project exploring the social and cultural construction of wellbeing in Bangladesh, Ethiopia, Peru and Thailand (McGregor 2007), and Young Lives, an innovative long-term international research project investigating the changing nature of child poverty in India, Ethiopia, Peru and Vietnam (Young Lives 2008). The paper concludes with some methodological reflections, including a brief summary of the challenges of different approaches.

1.1 Research on Wellbeing in Developing Countries

Research into wellbeing and subjective experiences in developing countries is growing rapidly,2 and represents a paradigm shift towards holistic, person-centred, and dynamic understandings of people’s lives, which are nonetheless embedded in particular socio-cultural contexts (Boyden 2006; Gough et al. 2007; Camfield, Streuli et al. 2008). People’s values, aspirations, and experiences of happiness or unhappiness are now measured directly within some large surveys of individuals and households (e.g. the World Values Survey, South African Quality of Life Trends Study) rather than inferred from proxies such as income (see Graham 2005; Guillen-Royo and Velasco 2009 for reviews of ‘happiness economics’ in developing countries). They are also explored in participatory studies that attempt to identify the pathways to particular outcomes (for example, chronic poverty), within the constraints of a cross-sectional study, and the perceived possibilities for change in the future (see Table 4, Appendix 1 for a summary of the key studies). The inclusion of subjective experiences and meanings is part of a move within international development and research on poverty from economic to multi-dimensional understandings of people’s lives.3 This is because the latter improves our understanding of the former (for example, the role of social norms and relationships in maintaining community savings groups in Bangladesh [Goetz and Sen Gupta 1996]).

However, even multidimensional understandings of people’s lives still focus on what they should have or be able to do, rather than what people think and feel about what they have and do (McGregor 2007), which is an obvious role for subjective measures and qualitative methods. Multidimensional approaches also fail to acknowledge the interpersonal and recursive aspects of wellbeing; for example, the “base” of shared social and cultural capital that defines “locally defined necessities or what a person needs to participate and aspire” (Gudeman 2004, p. 9). By this we mean that people’s experiences and evaluations of their lives are shaped by their perception of their environment and themselves, in the context of what they value and aspire to. The rationale for taking wellbeing as the focus is that it places the person, in their relationships and surroundings, at the centre; Rojas describes this as engaging with a “person of flesh and blood in her circumstance […] rather than the wellbeing of an academically constructed agent” [2007, p. 261]). Qualitative approaches in particular foreground the presence of both the respondent and the researcher, which highlights the fallibility of all data collection by emphasizing their role in its ‘co-creation’, and encourages reflexivity about the selective process of interpretation and representation.

Methods to explore wellbeing can be highly contextual, so people are not separated from their environments, and sensitive to human diversity and its interaction with the dynamics of power (McGregor 2007; Boyden 2006). The WeD group, for example, characterizes wellbeing as “a state of being with others, where human needs are met, where one can act meaningfully to pursue one’s goals, and where one enjoys a satisfactory quality of life” (2007). People’s hedonic experiences (for example, pleasure or emotions) are undeniably important: the New Economics Foundation defines wellbeing as the quality of people’s experience of their lives (Shah and Marks 2004) and the cognitive psychologist Kahneman characterizes it as wanting one’s current experience to continue (1999). However, people’s values and aspirations also play a role; in particular their response to the central question “is my life going well, according to the standards that I choose to use?”4 (Diener). White (2007) makes a further valuable distinction between living a good life (values and ideals), having a good life (material welfare and standards of living), and locating one’s life (experience and subjectivity).

As an integrative or ‘umbrella’ concept, wellbeing provides the opportunity to combine diverse literatures and methodological approaches to attain a holistic picture of people’s lives. It enables researchers to explore firstly the way that wellbeing is grounded in and mediated by social, cultural and political structures and processes. Secondly, the centrality of relationships to understanding, pursuing, and achieving wellbeing, and finally, the inevitability of trade-offs and contradictions in this pursuit. In the following section we briefly outline the role of qualitative and mixed methods approaches in researching wellbeing and provide examples of their value in studies with both adults and children.

1.2 Methodological Approaches to Wellbeing

We have loosely categorized the approaches to wellbeing discussed in this paper as qualitative and mixed methods, recognising that even ‘pure’ quantitative approaches often contain an element of qualitative work (for example, focus groups or interviews to generate the items for measures), aside from the ‘qualitative’ aspects of survey administration (Olsen and Morgan 2005). Similarly distinctions within the category of qualitative research relate as much to how researchers choose to position themselves as to any fundamental difference in ethos or method (for example, the claims made by some participatory researchers or ethnographers that their research empowers or provides unmediated access to people’s deepest experiences) (Table 1).
Table 1

Overview of the approaches described in this paper

Approach

Distinctive feature

Methods

Examples

Qualitative, including ethnographic and participatory approaches

Holistic & flexible, in-depth, usually arising from long-term engagement. More participatory work involves active, if often brief, engagement to involve people in describing, interpreting, and occasionally changing their reality

Interviews; narratives, e.g. biographical essays or accounts of illness; life histories, participant observation, predominantly task-based group or individual activities (e.g. body mapping, time use diaries)

E.g. understandings of a good life in Cairo (Wikan); concepts of wellbeing in Bolivia (Calestani 2008) and among the Cree (Adelson 2000); Consultations with the Poor (Narayan et al. 2000); studies of urban violence in Columbia & Guatemala (Moser 2003, 2004); developing a QoL toolbox in Madagascar (Farnworth 2004)

Mixed-methods

Combines the opportunities and challenges offered by different methods and analytical approaches; increases data validity through iteration & triangulation & enables targeting of findings to different audiences

Combination of survey, qualitative, & participatory methods or data (e.g. household surveys & participatory poverty assessments or life histories); analysis & presentation using quantitative & qualitative approaches

e.g. WeD, Young Lives, combining Participatory Poverty Assessment & surveys in Uganda (Lawson et al. 2006; McGee 2004) and Rwanda (Howe and Mckay 2007), combining surveys & life histories in Bangladesh (Baulch and Davis 2007), structured & open-ended exploration of people’s ‘visions of the good’ in South Africa (Clark 2002)

1.2.1 Qualitative

The sub-section on qualitative methods covers ethnography, and common ethnographic techniques such as semi-structured interviews and participant observation, and task-based ‘participatory’ methods such as drawing. It also discusses the benefits from taking a more participatory approach, whatever methods are used. Wellbeing’s location in the disciplinary and methodological borderlands makes it a classic object of anthropological attention (Thin 2005; Corsin-Jimenez 2007). Wilk suggests using ethnography to study ‘wellbeing’ in the same way as it is used to study ‘power’ (and by extension ‘poverty’) where

Unpacking the various meanings of this term has helped make many of us understand that it is objective and subjective, measurable and experiential aspects are really two parts of the same whole. The objective reality of the exercise of power cannot be separated from the beliefs and feelings that motivate, activate, and justify it. Victimisation and empowerment are subjective as well as objective”

(1999, p. 93)

Although some anthropologists accuse their discipline of neglecting wellbeing (e.g. Thin 2005; Wilk 2008), much early anthropology, for example, Margaret Mead’s study of teenage life in Samoa (1928), is grounded in a powerful critique of modernity, possibly driven by a romantic vision of the ‘other’ that relates to the anthropologist’s own desire to find an alternative vision of wellbeing (for example, one that draws on the deep connections to ancestors and community that were considered absent in the West). These motivations may be familiar to contemporary researchers of quality of life and wellbeing.

Despite applied anthropology’s checkered history (e.g. Keesing 1945), it has produced ethnography: a research methodology that is grounded and flexible, and capable of generating new and surprising information about the way in which people see the world (Hammersley and Atkinson 1995). This is even the case when “these ‘tacit understandings’ (Giddens 1997, p. 169) […are] so profoundly internalised that they cannot be asked about directly” (White and Petitt 2005, p. 26). For example, White and Petitt argue that using participatory methods to access local perceptions of the ‘good life’ may not “capture the deepest values of what people consider well-being” (for example, a respondent’s concern about the state of their eternal soul) as these are beyond the frame of a ‘wellbeing ranking’ (ibid). Thin similarly proposes that wellbeing “needs to be set in the broader context of anthropology’s concern with morality—with what it means to be good, to live a good life and to organize social processes and institutions that facilitate or inhibit virtue and well-being” (2005, pp. 4–5).

Ethnographic methods provide what the interpretive anthropologist Geertz characterizes as “thick description” which attempts as far as possible to provide an insiders’ perspective on people’s understandings and actions (“our own constructions of other people’s constructions of what they and their compatriots are up to” [1973, p. 9]). The illustration Geertz gives of thick description is the difference between a “blink” and a “wink”; one is an involuntary twitch, while the other can be a conspiratorial signal to a friend, or even a parody of that signal to confuse an observer. The physical movement is identical but each has a distinct meaning (“as anyone unfortunate enough to have had the first taken for the second knows” [Geertz 1973, p. 6]), which varies between different contexts and cultural systems. Geertz provides an example of thick description from his own research when he was an involuntary participant in a police raid on the village cockfight, which gave him an experiential understanding of one of the main components of Balinese wellbeing:

Getting caught, or almost caught, in a vice raid is perhaps not a very generalizable recipe for achieving that mysterious necessity of anthropological field work, rapport, but for me it worked very well. It led to a sudden and unusually complete acceptance into a society extremely difficult for outsiders to penetrate. It gave me the kind of immediate, inside view grasp of an aspect of “peasant mentality” that anthropologists not fortunate enough to flee headlong with their subjects from armed authorities normally do not get. And, perhaps most important of all […] it put me very quickly on to a combination of emotional explosion, status war, and philosophical drama of central significance to the society whose inner nature I desired to understand

Geertz (1973, p. 416)

Geertz’s description illustrates the potential of ethnography for understanding how people conceptualise wellbeing, how this changes over time and in response to particular experiences, and what sacrifices people are prepared to make in its pursuit.5 As people’s meanings are often deeply buried-Bourdieu describes how ‘‘what is essential goes without saying because it comes without saying’’ (1977, p. 167)—it may be necessary to supplement direct questions about wellbeing with participant observation, and listen to respondents speaking in their own terms, rather than in the slightly artificial context of an interview or participatory exercise. The examples given in the paper illustrate the potential of ethnography as a method for studying wellbeing; for example, exploring local understandings of wellbeing (e.g. Calestani 2008), collecting detailed information on the experiences of people living in poverty using ‘longitudinal ethnography’ (e.g. Wikan 1985), and contextualizing understandings of people’s lives and livelihoods (e.g. Reynolds 1991). While a rich and holistic understanding can be gained through ethnography or longitudinal qualitative research, reassuringly even a single interview can provide insight into what someone means by quality of life or wellbeing, as described in the section on mixed methods. This is especially so in the context of responding to a measure of wellbeing, as described in Section 2.2.

The conventional distinction between participatory and qualitative and ethnographic methods is an artificial one as participatory methods are obviously a subgroup of qualitative methods, and many qualitative and ethnographic studies of children’s wellbeing use group or task-based activities (e.g. Punch 1998). Ethnography can also encompass the altered researcher/‘researched’ relationship (for example, the anthropological classic “Reflections on fieldwork in Morocco” [Rabinow 1997]), where participants have some role in setting the agendas of the research, the course of data collection, and/or analysis and presentation of results. Conversely, so-called participatory methods encompass many degrees of participation, as illustrated by Hart’s ladder of child participation (1992), which moves from “child-initiated and directed” to “manipulation”. Participatory research on poverty is distinct from mainstream economic approaches in its emphasis on experiential aspects of poverty such as being respected, having meaningful choices, and being able to preserve one’s dignity (e.g. Brock 1999). As in all qualitative research, participatory studies aim to be ‘experience-near’, but they also aim to create a space for people to share and reflect upon their experiences and to conduct research that generates valuable outcomes for participants, policy makers, and practitioners. Good participatory work can ‘widen the lens’ to include aspects of people’s lives that are often overlooked in purely quantitative studies such as companionship, everyday pleasures, and sources of meaning that enable them to sustain their wellbeing in insecure and resource-poor environments (Laderchi 2001; White and Pettit 2005; Camfield and McGregor 2005). The process of collective discussion and decision is seen as a ‘good’ in itself, which can also spur collective action (e.g. the Participatory Learning and Action approaches advocated by Reason and Bradbury 2000), for example, the Children’s Forums in Vietnam carried out as part of Young Lives (Pham and Jones 2005).

Participatory research has become increasingly popular during the past twenty years, evidenced by the mainstreaming of Participatory Poverty Assessments in the 1990s (e.g. Norton et al. 2001) and the World Bank funded Consultations with the Poor study (Narayan and Walton 2000, 2002; Narayan et al. 2000), although its integration or ‘co-option’ into the mainstream has attracted some criticism (e.g. Cooke and Kothari 2001; White 1996). The main claims made by participatory methodologies is that they more closely reflect respondents’ worldviews than traditional, ‘scientific’ approaches by (i) recognizing the contextual, subjective and non-material dimensions of human experience, (ii) illustrating the complex dynamics behind poverty and well-being, and (iii) “draw[ing] out culture, location and social group-specific understandings of the dimensions of well-being” (White and Pettit 2005, p. 13). For example, White and Pettit (ibid) cite two volumes of practitioner ‘reflections’ on participatory methods, which note their value in identifying ‘improved quality of life according to local standards’ (Cornwall and Pratt 2002), and ‘capturing local perspectives’ (Cornwall et al. 2001).

Connecting with participants’ understandings is even more important for studies of children’s wellbeing (as explored in Sects. 2.3 and 2.4) because children’s interests and priorities may differ and even at times conflict with those of adults (Qvortrup 1994; Prout and James 1997; Woodhead and Faulkner 2008). Children’s participation in analysis as well as data gathering can increase the reliability of the research (Kirk and Miller 1986 and Kefalyew 1996, p. 204, both cited in Ben-Arieh 2005) and may also help diminish the ethical problem of “imbalanced power relationships between researcher and researched at the point of data collection and interpretation” (Morrow and Richards 1996, p. 100). Hill (1997) and Thomas and O’Kane (2000) provide several examples of how to involve children during the data collection process, for example, by selecting methods that enable them to control the form and content of the discussion, interviewing children on more than one occasion, working in small groups to aid collective interpretation, or having a few ‘peer analysts’ draw out important messages from other children’s accounts.

Participation is assumed to enhance children’s subjective wellbeing in the short-term through the act of participating in the research, and in the long-term as a means of improving the accuracy of data collected to inform child-related policy making. Doing research with children rather than on children enables researchers to engage critically with their own assumptions about children’s capabilities and about the nature of a good childhood, and create spaces where alternative voices can be heard. However, participatory research with children involves more than simply using participatory techniques as no method is inherently ‘participatory’ (White 1996; Morrow and Richards 1996) and research methodologies need to be flexible to communicate appropriately with different respondents and respond to changing contexts and emergent findings. Visual and interactive methods may be helpful, especially with younger children, to allow all participants to engage in generating and reflecting on their data (Crivello et al. 2008). Nevertheless, drawing, speaking confidently in front of a group, or talking with adults are culture specific skills which children acquire gradually.

1.2.2 Mixed Method

The sub-section on mixed methods distinguishes between different approaches to combining methods (for example, the timing of the qualitative input), and describes the role of qualitative techniques in drafting, piloting, and validating measures of wellbeing. Wellbeing is a truly “hybrid” concept because it is both objective and subjective and these aspects cannot be separated or reduced to each other (Wilk 1999). For this reason Wilk recommends mixing qualitative and quantitative methods as “any ‘indicator’ no matter how clever, is going to miss an essential quality of what needs to be measured. There is no alternative but to combine measurements with assessments of what the measures mean to the people being measured” (1999, p. 93). Both quantitative and qualitative methods provide “answers” on the nature and extent of wellbeing that can be used separately to validate each others findings (‘triangulation’) or brought together to deepen insights and possibly raise new questions. Mixed methods research6 is defined as “the combined use of both quantitative and qualitative methodologies within the same study in order to address a single research question” (Hewson in Jupp 2006, p. 179). However, this simple definition obscures the challenges involved in reconciling the different values, goals, epistemologies, and analytical approaches held by different disciplines (see Olsen 2006; Bevan 2007b). For example, Brannen describes how

Quantitative researchers have seen qualitative researchers as too context specific, their samples as unrepresentative and their claims about their work as unwarranted–that is judged from the vantage point of statistical generalisation. For their part qualitative researchers view quantitative research as overly simplistic, decontextualised, reductionist in terms of its generalisations, and failing to capture the meanings that actors attach to their lives and circumstances

(2005, p. 7, op. cit. Jones and Sumner 2008)

Disputes continue even within mixed methods research, for example, Bevan (2005) makes a further distinction between ‘Q-squared’ (called after the conference series of the same name, see www.q-squared.ca/) and ‘Q-integrated’ approaches as she contends that only the latter is genuinely integrated. According to Bevan, while ‘Q-squared’ approaches collect and analyse data separately and bring it together to answer specific research questions, ‘Q-integrated’ approaches entail cross-disciplinary research using a range of research instruments to produce different types of data that can be analyzed both quantitatively and qualitatively. An example of Q-integration is the research on subjective wellbeing by the WeD team in Ethiopia (Bevan 2007a, June) which reported aggregate responses to measures of wellbeing (Woodcock 2007) and also used individual responses in household case-studies alongside qualitative data (Lavers 2008; Pankhurst 2006).

Mixed methods approaches sequence the input from qualitative and quantitative methods in different ways, depending on whether they are combining methods to increase understanding, initiate new research questions, create complementary insights, or highlight contradictions for future exploration (Brannen 2005, pp. 12–14). For example, qualitative methods like semi-structured interviews, focus groups, and even ethnography (Ware et al. 2003) are often used to generate item content for measures of subjective health, which can be an important component of wellbeing. Cognitive interviews are used to understand more about how people respond to these measures (Barofsky 1996), both with established measures (e.g. Mallinson 2002), and as part of pre-testing (e.g. Bowden et al. 2002), and qualitative methods can be used in the validation of subjective measures, as they shift the focus from the measure to the respondent by aiming to assess the accuracy with which a measure has represented their worldview (Paterson and Britten 2003). Examples of cognitive interviewing and qualitative validation are given in Sect. 2.1.

2 Examples of Qualitative and Mixed Methods Approaches to Wellbeing

Section 2 illustrates the value of the approaches described above in exploring wellbeing with both adults and children. Sections 2.1 and 2.2 provides examples of studies with adults using qualitative (ethnographic and participatory methods) and mixed methods (‘participatory numbers’ (Chambers 2003), validating measures of wellbeing, and integrating qualitative and quantitative data). Sections 2.3 and 2.4 provides similar examples from studies with children using qualitative (ethnographic and participatory methods) and mixed methods (developing indicators or measures and integrating qualitative and quantitative data).

2.1 Studies with Adults Using Qualitative Methods

In this section we use Wikan’s longitudinal ethnography of a poor community in Cairo as an example of an ethnographic approach to wellbeing. This method has a long history as Oscar Lewis conducted similar studies with families in rural Mexico (1959) and Puerto-Rican families in New York (1966), which informed his theory on the “culture of poverty” (1966). Wikan (1985) carried out fieldwork with 17 families (100 individuals) from a small neighborhood in Cairo, which she visited regularly for 13 years (1969–1982), four of which were spent living with one of the participating families. Her sample was selected through “natural networks of friendship and amity” to help her develop a rapport and observe her respondents in their natural surroundings. Her main finding was that while the “expert judgment” is that standards of living are in decline,

In terms of their own values the poor argue that they dress better, have better health, better entertainment, more and better education, more employment opportunities, better chances for savings and investments, less need than before for borrowing, and much, much more in the way of prestigious consumer goods. And they celebrate the feast [Eid]

(1985, p. 8)

Wikan notes the effects on people’s wellbeing of a change in their self conception from seeing themselves as poor and shameful to experiencing increased self-esteem from having a better “style of living”. This expresses itself in new clothes, which enable nearly all families to celebrate the feast days, household equipment, and most importantly television. Television not only provides a focal point by attracting husbands and sons who previously hung out in cafes, but is also pleasurable as “the entertainment introduced in such lives by Egyptian TV comedy and soap opera genuinely changes the level of living” (1985, p. 12). She maintains that “as goods have accumulated, so have all family members self-pride. Whereas previously men were too ashamed of their homes to bring acquaintances there, they now have items to show off with pride” (1985, p. 24). Wikan assert that the importance of appearances in Egyptian society mean that material acquisitions play a significant role in increasing self-esteem and argues therefore for their inclusion in international measures of standards of living (noting also that standard measures such as housing have little relevance in this context, due to the tenure system and positive attitudes towards overcrowding). In addition to providing an alternative perspective on its constituents, Wikan gives a detailed account of people’s dogged pursuit of wellbeing (“mak[ing] swift use of marginal improvements in opportunities and circumstances” (1985, p. 23), and unpacks issues of intra-household allocation (for example, how employed daughters are more beneficial than sons for their mother’s wellbeing because they feel obliged to share their income).

Corsin-Jimenez (2007, pp. 1–2), in his edited collection on the anthropology of wellbeing, uses Evans-Pritchard’s evocative description of the Nuer of the southern Sudan as an illustration of how ethnography can contribute to our understanding of wellbeing through “extraordinarily rich descriptions of the social and political forms of life […which] provide an alternative route into the political and theoretical imagination of wellbeing”. For example, Evans-Pritchard defines wellbeing for the Nuer as “that in which a family possesses several lactating cows, for then the children are well nourished and there is a surplus that can be devoted to cheese making and entertaining guests” (1940, p. 21, in Corsin-Jimenez, op. cit.). Ethnographies can overcome a common criticism of wellbeing research as individualistic and politically naïve (e.g. Sointu 2005; James 2007) by providing socially and politically embedded accounts of wellbeing. James (2007, pp. 20–21) criticizes the current rhetoric of wellbeing for “a fastidiously modern and a historical presumption about how individuals ought to fare in life” and suggests that it is ‘absurd’ to look for wellbeing in contexts such as Sudanese refugee camps in Ethiopia where “a community’s historical sense of purpose has been evacuated—where people are told and slowly come to realize, that they will never return to their old ways of livelihoods”. Calestani (2008, this issue) similarly explores potential contradictions between individual and collective definitions of the good life in the Bolivian plateau, and Adelson (2000) notes how for the Whapmagoosti Cree wellbeing or “being alive well” is inextricably linked to the life of their community (“if the land is not healthy then how can we be?” [2000, p. 3]). Contributors to Corsin-Jimenez’s edited collection (2007) continue to emphasise the social and political dimension of wellbeing by tackling subjects such as the depoliticizing effect of international health and literacy programs in Nepal (Harper and Maddox) and changing concepts of wellbeing (mad ife) among the Fuyuge in Papua New Guinea in response to increased pressure from international mining companies and a modernizing state (Hirsch).

The interaction between global and local dimensions of wellbeing was also captured by Farnsworth’s work with smallholder organic farmers in Madagascar (2004), which aimed to develop a simple and flexible ‘QoL toolkit’ for use by local researchers. The methods needed to be specific enough to produce unique meanings in particular situations, and universal enough to speak to other stakeholders (for example, the German consumers at the other end of the organic supply chain). She also felt the methods should be dynamic as well as grounded as wellbeing is a process of becoming rather than a state. Farnsworth’s approach has obvious value as a way of fully understanding local people’s realities before developing a measure, and is quicker than the long-term ethnographic engagement described in the earlier part of this section.

2.2 Studies with Adults Using Mixed Methods

More recently participatory research has been used in developing countries in conjunction with household surveys (for example, Participatory Poverty Assessments), or as a means of generating quantitative data (e.g. Barahona and Levy 2003). Community wealth and more recently ‘wellbeing rankings’ are often used in the initial stages of participatory research as they provide valuable insights into how local people define wellbeing and the referents they use when asked to make judgments about their own lives (e.g. Seeley et al. 1995; Chadwick et al. 1995). According Van Campenhout (2007), the advantage of wellbeing ranking over conventional poverty indicators in developing countries is that they (i) use local conceptions of poverty and wellbeing that acknowledge the complexity of rural settings (Scoones et al. 1995), (ii) include intangible elements such as status, bargaining power, and access to support networks, and (iii) are less vulnerable to individual biases (for example, underestimating current consumption to increase the chance of receiving development assistance in the future). They also acknowledge inequalities in household distribution that affect the wellbeing of household members; for example, Van Campenhout noted that the major contribution of girls and women in fetching firewood and water were captured by participatory methods, but not by poverty indicators.

Findings from qualitative studies emphasise the importance of paying attention to people’s subjective wellbeing or QoL, and qualitative methods have an additional role in interrogating these constructs and improving their measurement. Within WeD subjective QoL was explored both qualitatively (Camfield 2006) and quantitatively (e.g. Woodcock et al. 2007). The WeD definition of QoL as “the outcome of the gap between people’s goals and perceived resources, in the context of their environment, culture, values, and experiences” (Camfield et al. 2006) was developed through a combination of literature review and exploratory research in the WeD sites7 to identify dimensions of QoL and test a range of methods. The definition builds on the World Health Organisation’s work on cross-cultural quality of life measurement (WHOQOL group 1995) by highlighting the interplay between people’s conceptions of their goals and satisfaction with their achievement, given their material and social circumstances. It shows the additional influence of ‘gap’ theories such as Calman (1984), Michalos (1985), and Ruta et al. (1994). During the exploratory research, WeD used the Global Person Generated Index8 (GPGI) as this represented the best operationalisation of the WeD definition. However, WeD subsequently developed an individualized measure of weighted goal attainment (the WeDQoL) where the level of satisfaction with a goal reported by the respondent is weighted by how important they perceived this goal to be. This is assumed to be a more accurate proxy for subjective wellbeing. The WeDQoL bridges the gap between the ideographic approach of the GPGI and nomothetic (abstract or universal) approach of international measures such as the WHOQOL. By developing and validating a questionnaire with a common format and additional items that reflected the priorities of people in particular countries (for example, having ‘metta-karuna’ for others in Thailand; see http://www.bath.ac.uk/econ-dev/wellbeing/research/methods-toobox/qol-toolbox.htm), it was possible to not only integrate subjective and objective data (for example, the Index of Needs Deprivation and total household expenditure, Camfield and Guillen-Royo 2009), but also to relate the subjective data to qualitative case study material for both individuals (Lavers 2008) and groups (Camfield, Guillen-Royo et al. 2008).

Cognitive interviewing is rarely part of the process of developing a subjective measure in developing countries, although arguably even more important due to the acknowledged problems of these measures outside their original context (Camfield 2004). Bowden and Fox-Rushby’s development of the KENQOL (Bowden et al. 2001, 2002; Fox-Rushby et al. 2003, Fox-Rushby and Bowden 2003; Nzioka et al. 2001) is an example of good practice as it involved first identifying the local concepts of health that the scale was based upon through extensive qualitative research and long-term participant observation, and then rigorously pre-testing the measure using six separate methods to ensure that it was capturing “the self-perceived and locally defined ‘health’ of people in Makueni district, Kenya” (Bowden et al. 2002). Qualitative methods can also be used as part of validating a measure, for example, Camfield and Ruta (2007) compared the content of the GPGI (Ruta 1998; Ruta et al. 1994) and in-depth semi-structured interviews as part of the validation of the versions that were administered in Ethiopia, Bangladesh, and Thailand during the WeD exploratory quality of life research (see also Martin 2007). This was a disconcerting exercise as more than half of the cases analysed found minor discrepancies between the GPGI and the interview, which related to the overall score, the relative weights in one area, or basic errors of comprehension. Some discrepancies can be attributed to interviewer competence; for example, Bangladesh had almost twice as many successfully completed GPGIs as Thailand. However, there were some patterns: in Bangladesh the discrepancies related to areas that were abstract, or personal, and thus difficult to capture in a few words (e.g. “own boredom and lack of fulfillment”) while in Thailand they mainly related to debt. This suggests that people will talk about different things in the more relaxed context of a semi-structured interview, not merely topics that are abstract, or idiosyncratic (i.e. important to them, but not ‘important’), but also ones that are potentially shameful. The impression is reinforced in Table 2, which compares the GPGI scores for importance and satisfaction given by a 73 year old man from rural Ethiopia with his responses in the accompanying semi-structured interview. His overall score was low but positive (63 percent), however, his description of his current situation was “I am living a dead life, and I want to die […] I am living a life that is horrible and very bad/worst”, which suggests a much lower score. Similar discrepancies can be seen in Table 2 between the weights given to areas and the number of times they were mentioned, and between the score for satisfaction and his verbal evaluations. This type of exercise underlines the need for caution in interpreting both quantitative and qualitative data outside the contexts in which they were collected. Since the respondent also said he wanted to die because he had no farming implements, this may be a manner of speaking that relates to his age, location, or religion, and would mean something quite different in the mouth of another respondent. It highlights the value of thick description, or at least having local anthropologists or historians at hand to resolve these puzzles (Table 2).
Table 2

Comparison of quantitative and qualitative data on the subjective experiences of an Ethiopian respondent

GPGI area

Quantitative data

Qualitative data

Wealth, poverty, & assets

Score 2 of 6, 30% weight, mentioned >20 times

It is good to die rather than to live in poverty. I am very poor

Education

Score 4 out of 6, 30% weight, mentioned indirectly twice

I am not skilful, knowledgeable, and sociable because of my poor living condition

Labour/work

Score 4 of 6, 20% weight, mentioned 16 times

I am not benefiting from my life because no return from my work

Health

Score 6 out of 6, 10% weight

I am getting physically weak and old

Peace

Score 6 out of 6, 10% weight, mentioned 5 times, but not in relation to the current regime

This day I am not happy with any thing but during the Haile Selassie regime I was happy with life

2.3 Studies with Children Using Qualitative Methods

Ethnographic approaches have been used in studies of childhood wellbeing, with the obvious proviso that it is less easy for an adult researcher to be a participant observer, or ‘blend into the background’. For example, Punch (1998) used a mixture of semi-participant observation, informal interviews and task-based methods to investigate how children negotiate their independence as they grow up in rural Bolivia. Corsaro and Molinari (2005) conducted a seven-year longitudinal ethnography of a group of Italian children’s transition from pre-school to middle school, which sets the unique opportunities offered by the Northern Italian school system in an international context and reflects on the potential threat from Italy’s changing political climate. Nonetheless, the authors are careful not to diminish children’s agency, and show how peer-to-peer social interactions are as salient to their educational development as support from parents and teachers, and create effective ‘rites of passage’ that maintain children’s wellbeing. Reynolds (1991) undertook a detailed study of children’s time use in the Middle Zambezi Valley that used case studies of precarious rural livelihoods to illustrate the impact of large-scale development projects such as dams and game parks on people’s well-being. By intensively studying the activities of 24 children her research provides an excellent example of how “small facts speak to big issues” (Geertz 1973, p. 23), namely the impact of market-driven development strategies on the wellbeing of rural populations. It is also a good example of mixing methods as interviews and observations were combined with extracts from children’s diaries and secondary data from district records.

There are many good examples of international participatory studies9 (e.g. Ungar 2003), which address the related concept of childhood resilience (often defined as maintaining subjective and psychosocial wellbeing in adverse circumstances, e.g. Masten 2001). For example:
  • Day-in-a-life (Gillen et al. 2007, http://dayinthelife.open.ac.uk/index.cfm), which filmed five days in the lives of 30-month olds in Peru, Italy, Canada, Thailand, and the UK to explore cultural differences in development and learning in early childhood

  • International Resilience Project (Ungar and Liebenberg 2005, www.resilienceproject.org), which examined how young people ‘grow up well’ in 14 challenging environments, despite exposure to what local informants characterized as atypical levels of risk, using culturally appropriate methods such as ‘talking circles’, and

  • Negotiating resilience study, which applied the day-in-a-life methodology to the lives of 13–15 year olds in matched sites in Canada, China, India, South Africa and Thailand who are making successful transitions between two (and possibly more) culturally distinct worlds (Didowsky, pers. comm.)

All these studies seek relatively unmediated access to children’s perspectives and experiences (for example, by using video diaries), and involve children and/or significant adults from their families and communities in interpreting the data. For example, teenage participants in Negotiating Resilience are asked to discuss both their own video tape and a tape from a child of the same gender in a paired research site and reflect on any similarities or differences (e.g. Saskatoon in Canada and Cape Town in South Africa were paired as both contain young people in ethnically-based informal settlements). This encourages participants to interrogate their own ideas about what constitutes wellbeing, in the context of examples from another country where similar challenges are responded to in a very different way.

As discussed earlier in the paper, a key finding from participatory research with adults in developing countries is that the quality of interactions and relationships matter as much to people’s wellbeing as the quality of their assets. This is especially true of family and community relationships which are both intrinsically and instrumentally valuable (Devine et al. 2007) and have positive and negative dimensions (Wood 2003). Johnston (2006) generated similar data with children when she used participatory ‘poverty trees’ in a collective exploration of children’s ideas about the causes (‘roots’) and outcomes (‘fruits’) of poverty. For example, Johnston found that the quality of family relationships featured in children’s definitions of poverty, and having parents who were absent or still very young was identified as one of the causes of poverty. Incidentally, the similarity of responses to questions about poverty and wellbeing in developing countries seems to support Jones and Sumner’s premise that “the distinction between the two concepts is perhaps overdrawn” (2008). It suggests that the concepts may be linked in the minds of both researchers and respondents through a “positive/negatively connoted dualism [of] well-being/poverty” (Neff and Olsen 2007, p. 12), in the same way that when people are asked to describe the experience of wellbeing, they invariably list the resources required to attain it.

Fattore et al. (2007) also used group dialogues, combined with individual interviews and self-directed “task-oriented projects” (for example, keeping a visual journal) to understand what Australian children (aged eight to 15) saw as “positive well-being”. The goal of the study was to identify new or important indicators to monitor the well-being of Australian children, for example, feeling valued and secure in relationships, being a ‘moral actor’ in relation to oneself and others, and being able to make choices and exert influence in everyday situations. Young Lives similarly used participatory methods to explore children’s wellbeing (Crivello et al. 2008) through group exercises and individual interviews with children and adults, which built on their previous participation in the group exercise. The exercises produced voluminous data on important components of wellbeing during piloting, especially from older children and caregivers, and showed surprising regularity across respondents and contexts (for example, community focus groups vs. individual interviews). Key themes that emerged in all countries were the importance of family support, education and recreation, good social relationships, and good behavior. The inclusion of good behavior may relate both to its role in facilitating smooth social relationships, and problems in the translation of wellbeing, which often gave it a strong moral tone. Common indicators of illlbeing were also predominantly social and respondents even described material indicators such as dirty clothes or irregular meals as reflecting a lack of care and support (Table 3).
Table 3

Summary of findings from Young Lives Country Pilot reports, Summer 2007

 

Ethiopia

India

Peru

Vietnam

Child wellbeing

Good harvest, fulfilling basic needs

Education & recreation

Health

Supportive family

Good social relationships

Religion & good behaviour

Household not poor (father has good job) & owns books & consumer goods

Have everything they need & eat good food

Has neatly combed and plaited hair, clean clothes and person

Happy, joyful, always smiling & pleasant, not shy or guilty

Bright & physically active, healthy & good looking

First class student

Mixes well with others, & doesn’t fight; generous to others & not snobbish

Family, love & support of parents

Education, incl. materials for study & recreation

Well-kept & fashionable clothes

Good behaviour

Nice house with TV

Working to support family alongside education

Health

Good local services

Consumer goods, motorbike, money

Studying abroad or teaching in the future

Child illbeing

Poverty & unfulfilled needs

Not attending school & starting work young

Orphaned or separated from parents

Family conflict

Disobedience

Family illness

Plays in the drains, kills birds, steals & has a ‘bad name’

Appears ‘dull’ & is too shy to talk to people, always weeping

Doesn’t bathe (or only once a week) & smells of urine

Looks dirty with uncombed hair & torn clothes

Father doesn’t support the household so mother has to work

Household doesn’t have money, sufficient clothes, or a television,

Father drinks & beats mother, family members always fighting

Behind at school & can’t read well

Doesn’t eat regular meals, unhealthy & frequently ill

Drugs, alcohol & gangs

Poverty

Lack of basic services

Poor food

Beatings from parents

Being an orphan or neglected by parents

Sexual abuse (girls)

Supernatural beings

Education (beatings, difficult subjects, studying hard)

Quarrelling with family & friends

Neglect by parents

Flooding & rubbish in neighbourhood

The data in the table is drawn from activities with children aged between 5 and 7 and 11 and 13 and focus groups with parents and other community members conducted during the pilot (Spring 2007)

As an illustration of the potential of this approach, we describe a wellbeing exercise conducted with boys aged 11–13 in Addis Ababa, Ethiopia (October, 2007). The example underlines the importance of comprehensive note taking as the most surprising insights came during the discussions of each child’s presentation (described below). According to the researcher’s notes, the first child to present (a 12 year old boy) was an orphan. He emphasised that “a child that is doing well has both parents. He has a house with many rooms, CD [player], and TV. He has a good variety of food prepared for him by his parents. The child goes to entertaining places with his parents. He goes to a school that has a field and equipment for kids to play on such as a shertete (slide), jiwajiwe (swing), and merry-go-round. The school is not far [from his home], it has good classrooms and clean toilets for boys and girls separately; and it also has a library”. The presenter characterised a boy who is not doing well as “having no parents and living alone. The roof of his house has holes so during the rainy season, water goes into the house and as a result the boy gets sad and cries. He doesn’t go to school and does not have any food to eat because his parents are dead”.

The other participants raised a number of questions about the boy who was doing badly, for example, why isn’t he helped by relatives or neighbours? (answer: “people do not get close to him because he has dirty clothes”), why can’t he do paid work such as shoe shining? (answer: “there is no-one to buy the boy polish for the shoe shining”), why can’t he get help from an NGO? [Non Governmental Organisation] (answer: “no-one gets close to him so he doesn’t have any access […] no-one can prove his problems to the Kebele [local authority] or NGOs). One participant observed that the child who was doing well didn’t have a school bag to carry his books, which seemed incongruous, but the presenter responded “does living well means being rich? No, living well does not mean being rich”. The two most important indicators for wellbeing ranked by the participants were getting a good education, because “‘education is key to achieving wellbeing”, and “having a good family that can advise the children”. “Getting a balanced diet” was only slightly less important because “if a boy does not get a balanced diet he would not understand what he learns”. The four indicators of illbeing generated during this exercise (being an orphan, lacking family support or proper “follow-up”, leaving school, and bad behaviour) were considered equally important and interlinked, for example, leaving school led to bad behaviour as “a child who does not learn will finally be a thief”. The participants were also asked how the situation of the child who had the worst life could be improved. Apart from one mention of basic needs, their responses centred around relationships (advice and moral education, receiving care and support from family, having positive role models and avoiding ‘bad boys’, and good relationships with family and neighbours) and the child’s own agency (studying and working hard, being obedient, sensitive to others needs, and disciplined) (Wellbeing exercise, October 2007).

2.4 Studies with Children Using Mixed Methods

Developing wellbeing indicators based on children’s experiences and perspectives is beneficial from an analytical as well as an ethical perspective as children are usually the best source of information on their daily activities (Ben-Arieh 2005). They can also provide reliable information on other aspects of their lives and children as young as seven can engage with abstract concepts such as ‘children’ and ‘human rights’ (Melton and Limber 1992). For this reason the Psychosocial Working Group (http://www.forcedmigration.org/psychosocial/PWGinfo.htm) used child-focused qualitative methods to develop a culturally appropriate measure of psychosocial wellbeing for use in post-conflict Afghanistan (2005). Davis et al. (2003) initiated the process by combining intensive participatory methods with children (the “Children’s Ideas Project”) and focus groups with parents and grandparents to learn how war-affected children in Kabul experience and understand their situations. Wellbeing was understood by respondents in four separate senses: as an ideal, as “hoped-for achievements”, as a standard for the important things in children’s lives, and as the qualities that children should develop (ibid, p7), and was centered on the local concept of Tarbia which refers to children’s manners and the quality of their relationships with others. A second study two years later (PWG 2005) developed and administered a 23-item questionnaire based on Davis et al.’s findings, which was used with children and adults to assess the effect on psychosocial wellbeing of three types of intervention (psychosocial, water, or a combination of psychosocial and water) and was combined with qualitative and participatory research and a sub-study on means of coping. The value of a mixed methods approach is illustrated by the fact that the quantitative and qualitative research presented contrasting results. Both considered the combined intervention the best, but the quantitative measure rated ‘water only’ as almost as effective as the combined and ‘psychosocial only’ as ineffective, while the qualitative results supported the value of both. For example, children said that the psychosocial intervention helped them communicate with parents and reduced beatings by teachers. The qualitative results also highlighted the gendered nature of risk and coping and enabled exploration of the differences between the sites identified in the quantitative results, which were hypothesized to relate to their internal cohesion and level of initiative in helping children.

Another example of combining qualitative and quantitative comes from Young Lives where econometric analyses of panel data from Ethiopia demonstrate that children who lost one or both parents early on are not only resilient, but may have better cognitive and educational outcomes than their peers (Himaz and Camfield 2009). This surprising finding challenges the homogeneity of the administrative category ‘orphans and other vulnerable children’ (Meintjes and Giese 2006) and draws attention to the importance of timing in predicting the effects of key events. Young Lives’ integrated data set enables the processes behind it to be explored with descriptive statistics and qualitative case studies using data from multiple sources (see Crivello et al. 2008).

3 Conclusion

While the authors’ perspective on mixing methods can be summarized as ‘what’s the alternative?’ it would be unwise to ignore the challenges this involves, especially on international collaborative projects. Some ‘methodological cautions’ with using qualitative methods in studying wellbeing in developing countries include their lack of credibility with certain audiences (for example, local policy makers) who may be more familiar with aggregate statistics. It can be difficult to find local researchers and translators with qualitative experience, due to the absence of qualitative research infrastructure in developing countries, and consequently data collection is costly. Even qualitative researchers who speak the local language/s cannot participate in every research interaction during a large scale project, which means interpreting some field data ‘second-hand’. This is a challenge even when the data is transcribed and accompanied with detailed field notes. A further potential loss of meaning occurs from working through an interpreter or with translated data, which means that is difficult to share qualitative data through data archives. Finally, qualitative approaches generate an enormous amount of data for analysis and even with qualitative data analysis software it is hard to share analysis across the team and ensure transparency and accountability in the conclusions drawn.

More participatory work presents further challenges and some authors have expressed discomfort with its recent entry into the mainstream, emphasizing that participatory methods can also be top-down and extractive. Other concerns are that the emphasis on community consensus rather than individual priorities may provide cultural context at the expense of individual experience, emphasise public goods over private (e.g. services rather than family relationships), or marginalize minority interests. There is often a pronounced ‘framing effect’ from its link to development practice, and starting from ‘poverty’ may miss the opportunity to understand people’s lives in their own terms. Finally, there are great variations in the quality of participatory research and the extent of participation that are not always apparent from the project reports. The challenges described in relation to qualitative methods also affect studies using mixed methods where researchers have to overcome unhelpful dichotomies (for example, between ‘hard’ and ‘soft’ data), ‘intellectual-stereotyping’, and disciplinary conflicts, which are often exacerbated by institutional structures. Finding researchers with a sufficiently broad skill-set is also challenging (especially in the study countries); as is the meaningful combination of data in analysis when the data collection was underpinned by different epistemologies. The need to evaluate and reconcile findings from different methods also necessitates a common means of establishing validity, which may require more inclusive criteria (see Sumner and Jones, this issue).

The examples provided in the paper demonstrate the value of qualitative research in its own terms, despite the above challenges, but have they answered the more pressing questions of what researchers more familiar with quantitative methods can learn from qualitative approaches, and how they can use them in their work? We advocate a pragmatic approach that engages on three levels:
  1. (i)

    with qualitative literature—ethnographies, social history, reportage, novels—to gain a fuller understanding of people’s contexts and influences;

     
  2. (ii)

    with the skills and knowledge of multi-disciplinary teams, which involves understanding and respecting the world view of other researchers as well as respondents; and

     
  3. (iii)

    with qualitative methods, which used strategically can both enrich and improve the accuracy of quantitative data.

     

For example, if a researcher wanted to adapt a measure of children’s subjective wellbeing for use in schools in Addis Ababa, a good place to start might be Poluha’s detailed ethnography of the Ethiopian school system (2004), or Tekola’s research on the experiences of children in Addis Ababa (2008), or findings from any of the cross-sectional and longitudinal studies that have worked in this area (e.g. Young Lives, WeD). The measure would need to reflect the different competencies and experiences of ‘school-aged children’ and the characteristics of the setting where it was administered. It would need thorough pre-testing to ensure the validity of the content and method of measurement, and it might be prudent to check conceptual validity (Herdman et al. 1998) through exploratory qualitative work. Cognitive debriefing would be helpful and might throw up unanticipated problems, which would be resolved through discussion with key informants and local or country-specific researchers who could set responses in their cultural context. Finally, when interpreting the data it might be useful to “complement it with other sources, especially qualitative sources, so that the dynamics, complexity, contradictions, and diversity of people’s positions can be understood in a nuanced way. It is likely that personal and family health, community involvement, and cultural capital may all contribute to well-being and that transport, insecurity, vulnerability and family worries may all contribute to ill-being […] but these are each in turn construed differently from place to place and from time to time” (Neff and Olsen 2007, p. 18).

Footnotes
1

Two of the three authors trained as anthropologists; two have also engaged extensively with psychological methods and literatures; and all three have conducted multi-disciplinary research with children and adults in developing countries.

 
2

See Gough and McGregor (2007) and studies by Biswas-Diener and Diener (2001, 2006), the International Wellbeing Group, Moller, Graham, Rojas, Camfield, etc.

 
3

See Sumner 2007 and the literature published by the WIDER research project on measuring human wellbeing (e.g. McGillivray 2006; McGillivray and Clarke 2006).

 
4

See also the work of philosophers writing on these issues such as Valerie Tiberius, Mark Chekola, and Daniel Haybron.

 
5

The ‘acceptance tale’ that Geertz recounts can also be seen as an example of the anthropological rhetoric critiqued by Clifford and Marcus (1986).

 
6

Cresswell (2003) and Brannen (2005), provide a useful overview of this area, and Carvalho and White (1997), Kanbur (2003), and Jones and Sumner (2008) explore the potential of these methods in relation to studies of poverty and children’s wellbeing respectively.

 
7

WeD’s exploratory research took place in rural, peri-urban, and urban sites in Bangladesh, Ethiopia, Thailand and Peru. The average sample size for the countries was 360 (range 314–419) and age and gender were used as the key breaking variables, followed by religion or ethnicity. The fieldwork used qualitative and quantitative methods, including semi-structured interviews, focus groups, the Global Person Generated Index and the Satisfaction with Life Scale, all of which had been piloted in similar WeD sites.

 
8

The Global Person Generated Index (GPGI) is an ‘individualised’ QoL measure that uses a mix of open-ended questions, scoring, and points allocation to establish people’s satisfaction with the areas of life that are most important to them. It was developed in 1994, revised four years later to broaden the focus from health-related QoL to QoL itself, and piloted in Ethiopia, Thailand, and Bangladesh in 2004 (Ruta et al. 1994, 2004; Ruta 1998)

 
9

See also Johnson et al. (1995) on environmental resources in Nepal; Woodhead (1998, 1999, 2001) on child labour; Ennew and Plateau (2004) on physical punishment; and Boyden and De Berry (2004) on reintegrating child combatants.

 

Acknowledgements

The authors thank the participants in WeD and Young Lives research, as well as the country researchers who generated much of the data referenced in this paper. Elaine Chase provided invaluable comments on an earlier draft of the paper. In relation to WeD, the support of the UK Economic and Social Research Council (ESRC) is gratefully acknowledged. Young Lives is funded by the UK Department for International Development (DFID) and based on a collaborative partnership between the University of Oxford, Save the Children UK, The Open University, UK, and a series of prominent national research and policy institutes in the four study countries.

Copyright information

© Springer Science+Business Media B.V. 2008