There are two elements that are crucial to building reliable evidence: (1) the existence of comparable treatment and control groups, and (2) the availability of measurable outcomes.
In randomized controlled trials (RCTs), a simple coin toss can be used to assign units to treatment and control groups, thereby creating comparable groups.
In encouragement designs, which are one form of RCT, individuals who are eligible for the intervention are randomly divided into an encouragement group and a control group, and only the encouragement group is actively encouraged to take up the intervention.
In cluster RCTs, the unit of randomization is groups—such as classes, schools, or municipalities—so that one can implement RCTs even when individual-level randomization is difficult.
Familiarity with statistical terms such as statistical significance and 95% confidence intervals makes it easier to understand the results of impact evaluations.
