Select a research design commensurate with the level of certainty required for your audience.

It is often help­ful to think about the cer­tain­ty state­ments a pro­gram would like to make about its impact. If a pro­gram is inter­est­ed in attribut­ing out­comes to pro­gram activ­i­ties (Cer­tain state­ments), then a rig­or­ous mon­i­tor­ing and eval­u­a­tion plan should be imple­ment­ed. This would like­ly require a greater lev­el of invest­ment than a plan that would attribute pro­gram impacts to anec­do­tal evi­dence (Cau­tious state­ments). Con­sid­er the types of state­ments the pri­ma­ry audi­ence will need to be informed, or con­vinced, by a mon­i­tor­ing and eval­u­a­tion plan. For instance, local stake­hold­ers may be sat­is­fied with anec­dotes about a pro­gram’s impact, while donors may require causal state­ments about a pro­gram’s impact. Please see Appen­dix G for exam­ples of research designs required to meet the three lev­els of cer­tain­ty described below.