According to Craig and Smyth (2004), when looking for evidence on the effectiveness of interventions, the process of random sampling provides the highest level of evidence, as
the differences between the intervention and the comparison groups are most likely due to the intervention and not to other factors.