Introduction. Participatory statistics: a 'win–win' for international development

Jeremy Holland

The practice and potential of participatory statistics in development research

Participatory statistics have gained a methodological foothold in the pluralistic world of development research. In recent years participatory research has established its credentials as an approach – with an accompanying set of tools – in which local people themselves generate statistics. Since the early 1990s there has been a 'quiet tide of innovation' (Chambers, 2008) in generating statistics using participatory methods, with diverse examples of cutting-edge and transformative participatory research that can be plotted in the NE quadrant of Figure 1.1. This tide has captured methodological innovation at all levels and in all spheres of development activity. Development practitioners are supporting and facilitating participatory statistics from community-level planning right up to sector- and national-level policy processes. Statistics are being generated in the design, monitoring and evaluation, and impact assessment of policies, programmes, and projects.

Reflecting on this accumulation of experience, this book suggests that a wider and more systematic use of participatory statistics would benefit both development agencies and local communities. The book makes the following claims for a 'win–win' perspective on participatory statistics:

• Participatory research can generate accurate and generalizable statistics in a timely, efficient (value for money), and effective way; and

• Participatory statistics empower local people in a sphere of research that has traditionally been highly extractive and externally controlled.

This book seeks to provide impetus for a step change in the adoption and mainstreaming of participatory statistics within international development practice. There is a wonderful opportunity for donors, partners, and development practitioners to reflect jointly on what an institutionalized approach to participatory statistics might look like and to agree a radical agenda for action. The challenge here is to foster institutional change on the back of the methodological breakthroughs and philosophical commitment described in this book. The prize is a 'win–win' outcome in which statistics are a part of an empowering process for local people and part of a real-time information flow for those aid agencies and government departments willing to generate statistics in new ways.

Participatory research is different from 'conventional' research

Participatory statistics are generated within a 'paradigm' of participatory research. This paradigm has long challenged a 'top-down' approach to knowledge generation that institutionalizes control of knowledge amongst powerful development professionals. Participatory approaches reposition ownership and control by asking 'whose reality counts?' (Chambers, 1997) and 'who counts reality?' (Estrella and Gaventa, 1998). In this way, participatory research respects local knowledge and facilitates local ownership and control of data generation and analysis.

In contrast to the individualized observation and discussions in much top-down investigation, participatory research also focuses on public and collective reflection and action. At its most political, participatory research is a process in which reflection is internalized and promotes raised political consciousness. In this way, population involvement in research shifts from passive to active. Participatory research supports empowerment by providing opportunities for local agency and shifting power dynamics in aid and development relationships (Eyben, 2006).

Participatory statistics come in many forms

Local people generate statistics in many ways, through mapping, measuring, estimating, valuing, and comparing, and combinations of these (Chambers, 2008, see Box 1.1). They do so through open-ended group-based data generation and analysis, accompanied by in-depth diagnostic or evaluative discussion. This is in contrast with conventional survey-based research, which is typically one-on-one and collects simple pieces of data using questionnaires with closed-ended questions (Barahona and Levy, 2003: 9). Crucially, when participatory numbers are compiled or aggregated, for example from a series of focus group discussions, they can be subjected to statistical analysis.

Powerful examples of counting are social and census maps (see, for example, Figure 1.2). Conducted in group mode, social mapping generates very accurate data in contexts where there is 'community owned' (public) knowledge, for example when listing and categorizing households in small rural communities. Through a process of 'group-visual synergy' (Chambers, 2008: 99), participants can 'see what is being said' and correct and add detail. For community census purposes, the outcomes have proven very accurate, and where there have been discrepancies, community analysts have wanted to check until they reach agreement.

An example of calculating comes from Community-Led Total Sanitation (CLTS), where as part of an appraisal process, local people worked out the quantities (e.g. cartloads for the whole community) of shit (the crude word is used) produced by their households in a day, multiplied out for longer periods, and added up for the whole community, concluding sometimes with community cartloads per annum (Kar, 2008).

Examples of participatory...