I am grateful to AGRODEP for selecting as me as one of the participants of the 2013 training workshop on "Impact Evaluation and Analysis of Development Intervention" in Dakar, Senegal and all the facilitators. I thank Alimatou Saadiya DIOP for her perfect logistic arrangements; Tanguy Bernard who taught us the importance of impact evaluation of development interventions and why it is crucial (at this stage of development) for African researchers to be more involved in this field of study going by statistics. My unreserved appreciation also goes to Jenny Aker who did so much within her limited timeframe to teach us sampling and power calculations; Alan de Brauw who taught us recent developments in the impact evaluation of Safety Net programs, with a particular focus on applying regression discontinuity designs to such; and Ruth Vargas Hill for exposing us to recent developments in the study of social insurance, with a particular focus on perfect and imperfect compliance in impact evaluation designs. I learnt a lot through the training on better ways to design and conduct research on impact evaluations as well as appropriate analyses that can be done against the crude method which I had used in the past (just looking for those who have adopted and those who have not adopted the intervention project/program and run a common regression such as probit or logit). I also learnt different approaches of randomization and how they can be used through explicit illustrations with case studies during the training workshop and different conditions when impact evaluation methodologies (such as Randomized Control Trial (RCT), Regression Discontinuity Design (RDD), Difference-in-Differences (DD), Propensity Score Matching (PSM), Instrumental Variables (IV)) can be used against the belief that any of these can be used as convenient for the researcher. Further, I learnt the appropriate conditions underpinning the use of different Treatment Effects (ATE, ATT, ITT, LATE) and how to estimate Intention to Treat (ITT), Local Average Treatment Effects (LATE) when we have perfect and imperfect compliance and when to use encouragement design. We were also exposed to social protection program design and evaluation; managing and minimizing threats to analysis when we have cases such as attrition and spillovers. Lastly, the workshop was an eye-opener on how to evaluate the impact of so many untapped researchable development interventions going on in areas of health and social protection, labour, agriculture, education existing in my country (Nigeria). In order to be more on impact evaluation, more trainings are still needed from AGRODEP and AGRODEP Impact Evaluation Network (AIEN) members.
