On this page:
- A variety of methods and approaches
- Frank Ellis: combining sample surveys and participatory techniques [eastern Africa]
- Khanya: the ‘vertical transect’ methodology [southern and central Africa]
- Livelihood trajectories [western Nepal, eastern India]
- Changing livelihoods: life histories and ‘cluster’ analysis [South Africa]
- Combining quantitative survey and ‘participatory’ research: Sechaba [Lesotho]
A variety of methods and approaches
Some examples of disparate and partly overlapping methods of studying livelihoods are briefly outlined here. In view of the prevailing emphasis of the Chronic Poverty Research Centre, they express a clear bias toward the study of rural livelihoods. It is important to recognise that this is not disconnected, however, either conceptually or methodologically, from the study of urban livelihoods along lines elaborated, for example, by Beall and Kanji (1999). There is a rapidly proliferating literature in this field, and - apart from the highly selective references singled out in the reseources section - the reader’s attention is also drawn to the following work that is specifically relevant:
- A series of country case-studies and comparative reviews produced under the auspices of the Institute of Development Studies (IDS) at the University of Sussex (for example, de Haan 1999, de Haan et al., 2000). [http://www.ids.ac.uk/eldis/ - click on 'sustainable livelihoods' under 'Research themes and weblinks']
- The Sustainable Livelihoods Working Paper series and the Natural Resource Perspectives briefings published by the Overseas Development Institute (ODI) [www.odi.org.uk/publications/working.php; http://www.odi.org.uk/nrp/index.php].
- The mass of country studies carried out in sub-Saharan Africa in the mid-1990s under the auspices of the De-Agrarianization and Rural Employment project (DARE) co-ordinated by Deborah Bryceson at the University of Leiden (for an overview, see Bryceson, 1999) [http://asc.leidenuniv.nl/research/dare.php].
- Elizabeth Francis’ valuable recent book Making a Living (2000), in which she explores the dynamics of struggle over livelihoods through comparative study of change in eastern and southern Africa.
This summary gives some indication of the complementary use of diverse methods of investigation in practice, on a relatively small scale, and also raises a question of general importance. Just as household livelihoods themselves straddle the boundaries between conventionally discrete economic sectors (industry/agriculture, formal employment/informal economic activity) and often the boundaries between conventionally discrete geographical spaces (urban/rural), so livelihoods research must transcend local ‘communities’ in order to comprehend both intra-household relationships and significant inter-household social relationships as these change over time. Both forms of relationship may be geographically ‘stretched’ over considerable physical distances. Neither form is readily susceptible to proper investigation through the conventional methods either of household survey work carried out within specified communities or of ‘participatory’ workshops confined to such communities. This immediately begs the question of an appropriate trade-off between work in one ‘community’ that purports in one way or another to be representative of that community, and work of a more dispersed but intensive kind that seeks to investigate at first hand disparate economic activities, and the relationships between them, that together comprise any one household livelihood but that often ‘stretch’ far beyond the physical boundaries of the community. For examples of such highly labour-intensive work, which inevitably gives weight to dense individual case-studies over the kinds of statistical generalisation that may be derived from sample surveys, see Murray (1995, 2000), who reports on investigations of changing livelihoods through the 1990s in the eastern Free State, South Africa. A vital part of the effectiveness of this kind of work was following up the same individuals and families at different points in time, scattered as they often were in different places, with repeated (small-scale) household surveys to plot demographic turnover and the trajectories of individuals’ experience. It could not, however, be claimed in any serious sense to be ‘participatory’.
1. Frank Ellis: combining sample surveys and participatory techniques [eastern Africa]
Frank Ellis, an agricultural economist at the University of East Anglia, is an important contributor to recent thinking on livelihoods, diversity and vulnerability, through influential articles (1998a, 1998b) and his recent book Rural Livelihoods and Diversity in Developing Countries (2000). In Part I of this book he develops in detail a similar framework of analysis to that outlined in the section above, and in Part III he elaborates the combination of survey and participatory methods that, in his view, opens up the possibility of a better understanding of rural livelihoods.
He outlines a critique of large-scale income surveys, on largely familiar grounds, with reference to studies undertaken in Ghana, Kenya and Tanzania in particular, and concludes that they are of very limited use in understanding changes in rural livelihoods over time. Smaller-scale sample surveys relating to particular communities or regional populations are of rather more use. He also advocates various different PRA methods (key informants, semi-structured interviews, informal group discussions, focus group discussions, Venn diagrams, etc.) ‘for discovering, quite quickly, the mediating processes within which livelihood strategies are adopted’, relating to social relations, institutions and organisations. He then offers a useful summary of the typical characteristics of different field methods, under the headings of large-scale sample surveys, small-scale sample surveys, semi-structured or participatory enquiry, and case-studies (Ellis, 2000: 196-7).
Ellis illustrates the practical application of these disparate methods through a case-study carried out in three villages in northern Tanzania in 1997. An experimental combination of methods was applied, guided by ‘considerations of cost-effectiveness and timeliness in obtaining policy-useful research results’ (Ellis, 2000: 200). They were:
- Semi-structured focus group discussions
- A participatory wealth-ranking exercise
- A sample survey of household demography and remittance income, farm incomes, non-farm income sources and household assets [30 households interviewed in each of three villages]
Typologies of livelihood strategies were derived from these exercises, and households were ‘mapped’ according to the relative robustness of income contributions derived from each mode of livelihood. With some qualifications and reservations, the expectation of the study was broadly confirmed: that focus group discussions and other participatory methods were well-suited to discovering the ‘vulnerability context’ of rural livelihoods; while sample survey methods were better suited to ‘examining more concretely how the assets and activities of the poor differ from those of the better-off’ (Ellis, 2000: 227).
2. Khanya: the ‘vertical transect’ methodology [southern and central Africa]
Khanya-managing rural change cc is a private consultancy based in Bloemfontein, South Africa, that in recent years has committed much of its work towards developing the applications of a livelihoods framework in the context of poverty reduction strategies. It has undertaken DfID-funded studies in Zambia, Zimbabwe and the Eastern Cape and Free State of South Africa, and on the basis of this experience drafted extensive ‘Guidelines for Undertaking a Regional/National Sustainable Rural Livelihoods Study’ (Khanya, 2000; see also Goldman et al., 2000a, 2000b). The Sustainable Livelihoods approach, in their view, does help to structure analysis of the support required to assist the poor. Particularly important are 1) a holistic analysis of strengths and 2) an understanding of macro-micro linkages. ‘Decentralised approaches are needed, in which the district level acts as the interface between micro-level understanding of clients, with macro-level policies, and provides the key intermediation in terms of matching poor people’s preferred outcomes and strategies with appropriate service delivery… At levels higher than district the complexity of achieving effective co-ordination and integration of services and programmes becomes too great’ (Goldman et al., 2000b, p. 4).
Khanya identified the advantages of the approach as follows (Goldman et al., 2000a, p. 3): its participatory methodology; the opportunity and need for pre-project ownership creation; developing a common methodology. Four levels were identified: the community level; the local service-provider level; the meso-level; the centre. Through their experience in three countries in southern Africa, they laid out a series of phases in developing a poverty reduction strategy using the Sustainable Livelihoods approach:
- Developing an interest in and commitment to such a strategy
- Organising a detailed study
- Conducting and writing up the study
- Developing the strategy
- Implementation of the strategy
The main political requirement was a sense of ‘ownership’ and commitment in appropriate departments of government at a fairly high administrative level. The principal methodological approach, however, was a variety of PRA exercises undertaken directly with local people. ‘The approach used builds from what we see (and don’t see) in the reality of people’s lives - so the focus of the study is on learning from people on the ground, the micro-level, understanding what impacts on their lives from that level, and how successive levels above support (or don’t) the operation at micro-level’ (Khanya, 2000, p. 5).
Khanya offers the experience of a vertical transect methodology, proceeding from an overview of policies at the centre to rapid assessments at village, district and provincial levels and then return to the centre. Such a vertical transect, they recommend, should involve an intensive study of about 6 to 8 weeks, with a multi-disciplinary team of 3 to 5 people for efficiency and economy. The team proposed the following more detailed schedule: a few days at the centre; one week on PRA investigations in case study locations; moving through layers of government service and support with workshops, semi-structured interviews and key informants; returning to case-study locations for at least one day to validate and triangulate, as well as further to explore specific identified issues; and arrangements for involving local people in the production of the report. Much weight was attached to the analysis of rural livelihoods in case-study areas. Some basic quantitative data could be gathered from participants, but the principal techniques were those associated with PRA, with an emphasis on ‘the community themselves’ defining their circumstances, needs and desired outcomes.
One problem with such an approach is that constituencies defined separately for the purpose of setting up focus groups, such as ‘farmers’, ‘women’, ‘youth’, ‘pensioners’ etc. often overlap in practice. The whole point of livelihoods research, after all, is to understand the ways in which diverse modes of livelihood are inter-related through the management of complex household portfolios in circumstances of structural change, not simply to identify the supposedly discrete concerns and interests of distinguishable social categories of the population.
3. Livelihood trajectories [western Nepal, eastern India]
[Bagchi et al., 1998]
The explicit research aim of the ESCOR-funded project on Long-Term Change and Livelihoods, initiated by the Overseas Development Group (ODG) at the University of East Anglia, was to describe and explain trajectories of change for individuals and groups over time, through fieldwork in 1996-7 in 15 villages in Western Nepal and two villages in West Bengal and Bihar. The comparative base for the work in Nepal was a sample survey of 667 rural households in Western Nepal carried out by the ODG in 1974-5. The team drew on other survey work undertaken in the mid-1970s to support the comparative longitudinal work in the other two areas.
Livelihood trajectories, according to the team:
- provide insights into the changing welfare and capabilities of individuals and of groups;
- can illuminate the process of change by revealing the ways in which negotiation, bargaining and struggle can alter circumstances;
- make it possible to ‘bridge the supposed micro-macro divide’ by a process of aggregation upwards from the lives of individuals;
- combine insights, in a seriously inter-disciplinary spirit, from the many different paradigms prevalent in development studies.
On the third point above, ‘progressive aggregation admittedly moves away from the rich and revealing, life-as-lived from day to day embeddedness of [Livelihood Trajectories], but increasingly makes possible generalisation and structural explanation as one moves upwards from households to local communities, districts, regions and states’. Livelihoods analysis ‘starts from daily lives and experiences’ but moves on ‘to explore not only "how people make history" but also the constraints that limit their functioning and capabilities’.
This project was explicitly ‘longitudinal’ in pursuit of its principal objectives, although not - for the most part - in the strict sense of following up the same individuals over time. Such research, while fraught with methodological problems, is potentially very valuable because of the time-scale of comparison. It is also highly unusual in practice, largely because of the massive wastage incurred through the familiar phenomenon of ‘loss of institutional memory’ over much shorter periods of time than the two decades considered here. A particular virtue of the article cited (Bagchi et al., 1998) is its explicit discussion of these methodological problems: for example, the difficulties of replication of the original sample(s) against the need for random sampling of population(s) 20 years later; the ‘gain’ of pursuing ‘original’ households against the ‘loss’ of a demographically unrepresentative range of households, etc. In the Nepalese case, the team found, only about 40 per cent of the original 1974-5 respondents were still living in 1996-97. However, without access to reliable baseline studies in West Bengal and Bihar that suited the particular purposes of the study, the team had to adopt ‘a more opportunist and eclectic construction’ of livelihood trajectories, through the use of various studies carried out at different times and for very different purposes.
The team used a variety of research techniques, such as village mapping, wealth-ranking, formal randomised sample household surveys and life histories. They also experienced major tension between quantitative and qualitative techniques. ‘The quantitative data provided the basis for showing what and emphasising what was representative, while the qualitative was able to reveal how and why and to highlight differences and variety within the range of human experiences in the areas studied - experiences that could help explain, problematize, and contextualize differences and changes in average values of variables from the quantitative survey’ (Bagchi et al., 1998: 461).
4. Changing livelihoods: life histories and ‘cluster’ analysis [South Africa]
[http://idpm.man.ac.uk/rsc/mlscza/methods.php]
The objectives of the ESCOR-funded Multiple Livelihoods and Social Change project (life-span 1998-2001) were as follows:
- to analyse socio-economic differentiation in two densely populated but relatively remote (former ‘homeland’) areas of South Africa;
- to identify the intervening socio-economic, political and institutional variables that affect efforts to alleviate poverty through improved livelihood opportunities;
- to develop a distinctive combination of methods for the study of multiple household livelihoods as they change over time, and for relating changes at the micro-level to changes at the macro-level.
Fieldwork was undertaken in the Central District of North-West Province and ‘greater’ Qwaqwa, Free State. Elizabeth Francis worked for 3 months in two adjacent villages in North-West in March-June 1999. She conducted forty-one life history interviews with people in forty different households in these two villages and with farmers on nearby state land. She used a unified interview framework that included questions about contemporary livelihoods. Interviews lasted around two hours. She also conducted interviews dealing with the local and regional institutional context with chiefs, headmen, local councillors, other local political activists, members of local community-based organisations, district council officials, provincial government departments and the National African Farmers’ Union.
Her sample was constructed in order to capture differences in livelihoods, resource access and income levels. She used snowball sampling, in order to understand inter-household relations through looking at clusters of associated individuals and households. Commercial farmers led her to people they had employed, members of other households introduced her to their kin and in-laws, to people they brewed beer with, people who herded their stock or helped them regularly, people who paid them to do domestic work, or to people who fostered their children. In following up relationships between individuals in different households, there was a tension between her wish to explore these further and her understanding of the importance of keeping the sample as widely-based as possible, in order to avoid potential biases. She asked informants about dispersed household members, kin with whom there was a lot of contact, and other people with whom there were relationships which impinged on livelihoods.
Rachel Slater carried out one year of fieldwork in ‘greater’ Qwaqwa between August 1998 and August 1999. She worked in ten research locations that were selected to represent different modes of livelihood, geographical locations, environmental conditions and settlement histories. The sequence of investigations is briefly summarised here.
- 29 interviews were carried out with traditional leaders and local and provincial government officials. These allowed identification of key informants who offered insights into historical and contemporary livelihoods issues in Qwaqwa and who facilitated the research process.
- 125 semi-structured interviews were carried out with a minimum of ten households in each location, to provide baseline information about household size and composition, sources of household livelihoods and patterns of migration in the pursuit of livelihoods. No rigid sampling pattern was used, but as broad a range of people as possible was included.
- Ten ‘clusters’ of households were identified, defined by significant relationships between members of different households who collaborated in different ways in the generation of their respective household livelihoods, and the connections were followed up in detail.
- 40 individuals were selected from the clusters and life history interviews were carried out, which offered a retrospective view on trajectories of change and facilitated an understanding of how livelihoods had changed over time.
By placing these changes in the context of local and national transformations, it was possible to relate some of the household-level changes to socio-economic, political and institutional changes at the regional, provincial and national levels.
5. Combining quantitative survey and ‘participatory’ research: Sechaba [Lesotho]
Sechaba Consultants are "a Lesotho-based, independent research company committed to presenting the socio-economic experience and concerns of local people in the development process". They have carried out a series of poverty studies in Lesotho through the 1990s, which now constitute a valuable longitudinal series (Sechaba Consultants, 1991, 1994, 1995, 2000), not in the strict or narrow sense of having followed the same ‘households’ over that time period but in the looser sense that similar or comparable questions were asked in 1991, 1994 and 1999-2000 to households sampled from the same areas. In the case of the most recent study, the survey was of 3,280 households in 130 villages and urban areas scattered across the country. The 1999-2000 work partly builds on their earlier work, in order to achieve comparability of conclusions concerning trends and patterns of poverty (Phase I), and also adopts a sustainable livelihoods model adapted from that of CARE International (see, for example, Mohasi and Turner, 1999, p. 15), which was explicitly committed to facilitating the expression of the views and priorities of Basotho themselves (Phase II). They reflect throughout the report (Turner et al., 2001) on the tensions which arose out of the combination of the two paradigms: a large-scale household questionnaire survey (Phase I) with a participatory livelihoods analysis (Phase II) reflecting the categories produced by and the experiences emphasised by respondents.
From the results of the large-scale household survey they created quantitative variables within the terms of reference of the livelihoods framework adopted, relating 1) to household assets, 2) to shocks and stresses, 3) to household activities or, in their terminology, livelihood choices, and 4) to household well-being. One recurrent source of tension, however, was that it was very difficult to apply local definitions of well-being in a standardised way for the country as a whole. They therefore applied two different definitions, and strove consistently to avoid confusion between the two.
The advantage of the methodology applied in Phase I was comparability of research findings with the 1991 and 1994 surveys. The 1999-2000 work confirmed, for example, that deep poverty continued to be concentrated in remote mountain areas, and also identified a new form of poverty prevalent amongst the poorest people in urban areas. Examination of macro-trends relating to the decline of formal employment both in South Africa and in Lesotho through the 1990s helped to interpret the survey findings relating to the declining proportion of households in all quintiles with a member in wage employment. The advantage of the methodology applied in Phase II was a very different emphasis – by comparison with prevalent diagnoses of Lesotho as a ‘basket-case’ requiring massive external intervention to relieve poverty – on the practical ingenuity and considerable achievements of ordinary people in striving to overcome their worsening socio-economic predicament. The analysis did reveal, however, an apparently disproportionate weight still attached to agriculture as a livelihood strategy; an inappropriate faith in job-creation as a route out of poverty; and a certainly misplaced attribution to ‘government’ of the responsibility for solving the problems of the poor. The ‘strategic vision’ developed by Sechaba placed emphasis for purposes of practical policy on two vital forms of support: the need to facilitate people’s own efforts to enhance their livelihoods, particularly in the western and northern lowlands and foothills; and the need to ensure improved safety-net provision, particularly in remote mountainous areas and for the urban poor. All policy interventions, however, had to be predicated on a "coordinated and effective response" to the massive crisis of HIV/AIDS (Turner et al., 2001, p. x).