We evaluate programs and polices designed to improve well-being in Africa.

Evaluations

Methodologies

We evaluate programs and policies designed to improve the well-being of Africans. Zerihun Associates works with government agencies, entrepreneurs, NGOs, donors, and other partners who are more interested in learning the answer to the question: How effective is our program? There are many methods of doing evaluations , the following are among the most commonly used methods by Zerihun Associates - includes both experimental and quasi-experimental methodologies.

Randomized Controlled Trials

Randomized Controlled Trials also known as Randomized Evaluations are part of a larger set of impact evaluations. Randomized Evaluations are deemed the gold standard of impact evaluation, because they consistently produce the most accurate and unbiased results. It is a type of impact evaluation that uses random assignment to allocate resources, run programs, or apply policies as part of the study design. Like all impact evaluations, the main purpose of randomized control trials is to quantify how large that impact is. Zerihun Associates have successfully managed over 30 randomized control trials across Africa. Click here to learn more about randomized control trials. Also, IPA and JPAL websites have a great deal of resources on RCTs.

Quasi-expermental Designs

If experimental evaluation technique, randomized evaluation, is not feasible, we work with clients to design the second best alternatives by using robust quasi-experimental techniques such as Propensity Score Matching and regression discontinuity design. PSM is regression-based and estimates “program effect” by eliminating observable confounders on which comparisons are made. If DD-PSM is available, then DD will take care of the effect of time invariant unobservable cofounders. Tests of significance based on a comparison of means (t-test) will be performed on the mean coverage values for variables of interest between intervention and control groups.

If treatment assignment is based on specific criteria, regression discontinuity design might be more appropriate to estimate the causal impact of the envisaged project. Using different competing estimation techniques will allow us to verify the robustness of our findings. We may also use different estimation techniques depending on the nature of the outcome variable