In 2014, we directed you to the 3ie database of impact evaluations. Here is another - the American Economic Association's registry for RCTs - one (implicit) purpose of which is perhaps to discourage ex post data mining with the focus on design and hypotheses. 29 of the 329 registered trials are tagged as agriculture, many of which are relevant to CGIAR (for instance, see this https://www.socialscienceregistry.org/trials/559 and this https://www.socialscienceregistry.org/trials/449).
The International Initiative for Impact Evaluation (3ie) has just launched a database of over 2500 impact evaluations. Of particular interest to the CGIAR community are the 249 evaluations tagged 'Agriculture and Rural Development', and an additional 60 'Environment and Disaster Management' and over 1550 'Health Nutrition and Population' IEs.
For instance, read this 3ie/IWMI evaluation of a electricity tariff policy change in West Bengal (India). JV Meenakshi (SPIA member) and co-authors found no impact of this policy change on total number of hours of groundwater pumped by pump-owner farmers, the number of hours of groundwater used for irrigation, or the number of hours of groundwater sold to other farmers. However, seasonally disaggregated results revealed a significant drop in the number of irrigation hours by owners of electric pumps during the summer. More here.
Relatedly, SPIA is commissioning work on CGIAR irrigation and management research to understand the comprehensiveness, strengths and limitations of these assessments under Activity 3.2 of SIAC.10Jun 2014