Spill-over effects refer to individuals in the control groups that are affected by the treatment in physical ways or in the form of price changes, learning or imitation effects. But individuals in the treatment group can also be affected by spill-overs, e.g. changes in migration patterns through being attracted by the treatment can have an effect on the impact of the programme. In the case of Mexico’s PROGRESA conditional cash transfer programme, spill-over effects caused by migration were detected, but the good news is that these spill-overs, if significant, can be measured and checked for. For example, the level of treatment exposure within groups can be adjusted to assess the magnitude of potential spill-over effects.

In addition to these technical challenges, potential ethical challenges should not be overlooked. The implementation of RCTs is not always feasible because of ethical considerations, e.g. how can it be justified that certain individuals are assigned to a treatment group while others are excluded from a potentially beneficial treatment.