A multidisciplinary (or better still an interdisciplinary) approach, which fully recognizes
the complexity of relationships between ‘‘variables’’ and of possible chains of causation,
particularly when considered over the medium and long term and when the ‘‘noise’’ of
external factors becomes greater;
(b) An exceptionally wide-ranging review—and careful analysis—of existing information
(objective and subjective; quantitative and qualitative) on development and change covering
an exceptionally long period of time;
(c) An effective integration and synthesis of the information collected, reviewed, and analyzed,
utilizing a strong common evaluation framework routed in systems thinking. This in order to
be able to generate preliminary hypotheses (narratives, suggested chains of causation) to be
tested for robustness and plausibility, and subjected to argument, debate, and the development
of counterhypotheses; and
(d) Independent ‘‘reality’’ checks of preliminary theories, including re-visiting the original
information reviewed, re-considering the review and analysis, and where appropriate
undertaking further research, including fieldwork. There is a need for qualitative evaluation
as interpretative lens and to humanize the statistics, recognizing the limits of quantitative
evaluations as adequate contribution to knowledge generation.