Tuesday, December 21, 2004
12.30 - 2.00 pm
Presenters: Michael Woolcock & Katherine Whiteside
Social development projects--most notably those using 'participatory' approaches to help select, implement, and maintain idiosyncratic local initiatives--seriously compromise experimental control over the variables under study and/or the observational environment. As such, they pose numerous challenges for project designers and confound many standard evaluation techniques. We critically examine these challenges in relation to two integral components of many social development programs, namely (a) service delivery that is both discretionary and transaction-intensive, and (b) localized control of resources and decision-making. Developing more credible and useful evaluations of social development projects requires explicit attention to internal assignment and adaptation processes, better measures of abstract social and institutional phenomena, tracking indirect impacts, and evaluating specific links in hypothesized causal pathways. Doing so will entail integrating data of varied types, sources, and quality, and incorporating different analytic methods and approaches to understanding causal inference. We offer brief empirical illustrations from international social development projects and from the U.S. experience with "community initiatives."
Materials from this event are offered in the highlight box to the right. Use the free Adobe Acrobat Reader to view the PDF files.
Back to Training Events and Materials.