Qualitative and Quantitative Poverty Appraisal: Complementarities, Tensions and the Way Forward*
Total Page:16
File Type:pdf, Size:1020Kb
Q-Squared - Qualitative and Quantitative Poverty Appraisal: Complementarities, Tensions and the Way Forward* Ravi Kanbur (Editor) Cornell University Email: [email protected] With Contributions by: Ravi Kanbur Robert Chambers Patti Petesch Norman Uphoff Martin Ravallion Francois Bourguignon David Sahn Caroline Moser Christopher Barrett David Booth Vijayendra Rao Luc Christiaensen Jesko Hentschel Paul Shaffer Rosemary McGee Ronald Herring Gary Fields Alex Wilks Erik Thorbecke Q-Squared Working Paper No. 1 October 2005 * This volume originally appeared as: Cornell University Applied Economics and Management Working Paper 2001-05, May, 2001. Q-squared • Centre For International Studies • University Of Toronto 1 Devonshire Place, Toronto ON M5S 3K7 Canada t: 416-885-7721 • f: 416-652-1678 • e: [email protected] CONTENTS Q-SQUARED? A COMMENTARY ON QUALITATIVE AND QUANTITATIVE POVERTY APPRAISAL RAVI KANBUR 1 QUALITATIVE AND QUANTITATIVE POVERTY APPRAISAL: THE STATE OF PLAY AND SOME QUESTIONS RAVI KANBUR 16 QUALITATIVE APPROACHES: SELF CRITICISM AND WHAT CAN BE GAINED FROM QUANTITATIVE APPROACHES ROBERT CHAMBERS 21 THE BEST OF BOTH WORLDS? ROBERT CHAMBERS 25 SELF-CRITICISM AND OBSERVATIONS ON THE WAY TO FINISHING VOICES OF THE POOR: FROM MANY LANDS PATTI PETESCH 29 BRIDGING QUANTITATIVE-QUALITATIVE DIFFERENCES IN POVERTY APPRAISAL: SELF-CRITICAL THOUGHTS ON QUALITATIVE APPROACHES NORMAN UPHOFF 32 CAN QUALITATIVE METHODS HELP QUANTITATIVE POVERTY MEASUREMENT? MARTIN RAVALLION 37 QUALITATIVE AND QUANTITATIVE APPROACHES TO POVERTY ANALYSIS: TWO PICTURES OF THE SAME MOUNTAIN? FRANCOIS BOURGUIGNON 43 STRENGTHENING QUANTITATIVE METHODS THROUGH INCORPORATING QUALITATIVE INFORMATION DAVID E. SAHN 46 ‘APT ILLUSTRATION’ OR ‘ANECDOTAL INFORMATION?’ CAN QUALITATIVE DATA BE REPRESENTATIVE OR ROBUST? CAROLINE MOSER 51 INTEGRATING QUALITATIVE AND QUANTITATIVE APPROACHES: LESSONS FROM THE PASTORAL RISK MANAGEMENT PROJECT CHRIS BARRETT 55 TOWARDS A BETTER COMBINATION OF THE QUANTITATIVE AND THE QUALITATIVE: SOME DESIGN ISSUES FROM PAKISTAN’S PARTICIPATORY POVERTY ASSESSMENT DAVID BOOTH 59 POTTERS AND SLUMS: TWO QUALITATIVE/QUANTITATIVE PROJECTS IN INDIA VIJAYENDRA RAO 64 THE QUAL-QUANT DEBATE WITHIN ITS EPISTEMOLOGICAL CONTEXT: SOME PRACTICAL IMPLICATIONS LUC CHRISTIAENSEN 69 INTEGRATING THE QUAL AND THE QUAN: WHEN AND WHY? JESKO HENTSCHEL74 DIFFICULTIES IN COMBINING INCOME/CONSUMPTION AND PARTICIPATORY APPROACHES TO POVERTY: ISSUES AND EXAMPLES PAUL SHAFFER 79 QUALITATIVE AND QUANTITATIVE POVERTY APPRAISAL WORKSHOP: SOME REFLECTIONS AND RESPONSES ROSEMARY MCGEE 84 i WHAT IS REQUIRED TO REDUCE TENSIONS AND INCREASE COMPLEMENTARITY? ROSEMARY MCGEE 88 DATA AS SOCIAL PRODUCT: PROBLEMS IN QUALITATIVE AND QUANTITATIVE WORK RON HERRING 90 THE EMPLOYMENT PROBLEM IN SOUTH AFRICA: FROM COINTEGRATION TO MR. ISAACS GARY S. FIELDS 94 POVERTY RESEARCH: EXTRACTIVE OR EMPOWERING? ALEX WILKS 98 TENSIONS, COMPLEMENTARITIES AND POSSIBLE CONVERGENCE BETWEEN THE QUALITATIVE AND QUANTITATIVE APPROACHES TO POVERTY ASSESSMENT ERIK THORBECKE 102 ii Q-SQUARED? A Commentary on Qualitative and Quantitative Poverty Appraisal Ravi Kanbur Cornell University April 2001 SUMMARY This compilation brings together the proceedings of a workshop on “Qualitative and Quantitative Poverty Appraisal: Complementarities, Tensions and the Way Forward.” Contributors were asked to submit short summaries of their positions, with detailed references to the literature as necessary. The compilation represents a remarkable statement of the state of the art and the debate on “Qual-Quant”, at a time when the complementarities between the qualitative and the quantitative traditions in poverty analysis are being recognized, but the tensions are ever present, and analysts and policy makers are looking for a way forward in using the two approaches to design effective poverty reduction strategies. The workshop spent a fair amount of time characterizing the two traditions along different dimensions, with (at least) the following five dimensions being identified: 1. Type of Information on Population: Non-Numerical to Numerical. 2. Type of Population Coverage: Specific to General. 3. Type of Population Involvement: Active to Passive. 4. Type of Inference Methodology: Inductive to Deductive. 5. Type of Disciplinary Framework: Broad Social Sciences to Neo-Classical Economics Different contributors collapsed these dimensions in different ways, and used different sets of terminology. But there seems agreement that points more to the “left” in the above spectra are more “qualitative” in nature, while those more to the “right” are “quantitative”. Some participants pronounced themselves to be in the “extreme center”. What is important, it was generally agreed, are the strengths and weaknesses of locating at different points along these dimensions, for the purpose at hand. It was recognized that there are indeed strengths and weaknesses at each end of a given spectrum. Numerical information can be more easily aggregated, but it can miss out on nuance and texture. General coverage aids representativeness, but can lose context. Q-squared • Centre For International Studies • University Of Toronto 1 Devonshire Place, Toronto ON M5S 3K7 Canada t: 416-885-7721 • f: 416-652-1678 • e: [email protected] Statistical inference can help in discussion of causality, but misses out on the power of inductive approaches. And so on. The key, then, is how to make the best of complementarities while minimizing tradeoffs. There was general support in the workshop for “small movements” from either extreme of any of the five dimensions. The support was strongest, almost universal, for the first two dimensions. For example, those in the qualitative tradition agreed that some numerical information could and should be collected in participatory poverty appraisal (PPA). There was also agreement (less strong) that the credibility of qualitative studies with policy makers and others would be greater if site selection could be put into a sampling frame – preferably, the same sampling frame as for a nationally representative household survey. The PPA exercise, for example, could be done on the site using its characteristic methodology as before, but the results could now be situated in a broader context. By the same token, those in the quantitative tradition agreed that the introduction of more subjective and open ended questions in a standard household survey could in fact provide a better handle on defining and measuring poverty even in standard income/consumption terms. Moreover, initial participatory exercises could suggest questions for inclusion in the standardized surveys. However, it was also recognized that there were limits to how much of such “simultaneous mixing” could be done without losing the essence and the strengths of either end of the spectrum in question. In any event, such mixing was more possible for some dimensions than others – the workshop did not have as many useful suggestions for mixing along dimensions 4 and 5 above, for example. In view of this, the issue arose of “sequential mixing”, where each end of the spectrum is called upon to do its best, and the integration takes place between the results of the two approaches so devised. It was recognized that this was difficult, not least because of disciplinary divides in dimension 5. Moreover, such mixing would require institutional arrangements to take account of the current dominance of the quantitative approach in policy-making circles. Relatedly, those in the qualitative tradition were more concerned about the ethics of doing research on poor populations with no direct benefit to the poor themselves (and considerable benefit to the researchers). They emphasized the importance of communicating research findings back to the poor, in a language they could understand. These issues were not on the agenda of those in the quantitative tradition, whose rationale is more that if research influences policy directly or indirectly and thus helps the poor in turn, then it is worthwhile. It was generally agreed that this first Qual-Quant workshop was hugely successful in what it did. But we have probably got as far as we can at this level of generality. An interesting dialogue has been opened up among some of the leading practitioners from both traditions. The next step should be to get more specific, focusing in greater detail on particular case studies of simultaneous or sequential mixing, to understand in greater depth the strengths and weaknesses of different points along the five dimensions identified above, and complementarities and tensions in mixing them. 2 I. INTRODUCTION Poverty analysts in the “Qualitative” and in the “Quantitative” traditions have been highly active in the policy debates of the past decade. While quantitative approaches have been dominant, especially in policy-making circles, the use of qualitative approaches has been increasing. Many bilateral and multilateral agencies now routinely commission studies in this tradition. There have also been increasing attempts at integrating the two approaches. But systematic attempts have been few, and successes have been even fewer. While there is a general acceptance, at least at the level of rhetoric, of the obvious complementarities between the two approaches, the tensions are more than apparent. The situation has undoubtedly improved compared to a decade ago, but practitioners in the two traditions still seem to inhabit unconnected worlds, with their own conferences, their own academic journals, and separate