Randomized Controlled Trials

The aim of establishing an evidence-based practice—in health care, psychology, or any other field—is to make clinical work, and the implementation of interventions, more scientific and empirically grounded. If this goal is achieved, care will be safer, more effective, and more consistent. Good research helps ensure that interventions are backed by evidence of sufficient quality to justify investment in implementing them, and later, scaling them up to benefit more people. Moreover, if policy makers propose to invest in an intervention (such as make a medication, a parenting program, or a school-feeding scheme available), then one of the central questions that they should ask is whether or not that intervention works. Does it achieve the outcomes that are expected of it, so that it will be a worthwhile investment of taxpayers’ money?

A randomized controlled trial (RCT) is essentially an experiment where some people receive a treatment (medicine or intervention) and another group (the control group) does not. The intention is to compare what would have happened if the intervention was not received with what happens when the intervention is received. RCTs are considered the gold standard of program evaluation, representing the best way to determine whether new interventions are effective or not. Evidence-based practice is built upon the foundation of the RCT, and it is rare, certainly in clinical practice, for evidence other than that generated by an RCT to be considered sufficient. Some, however, have argued that the dominance of the RCT marginalizes intervention types that do not lend themselves to an RCT design. Such arguments are important but beyond the scope of this entry. The sections that follow describe what an RCT is, show why randomization is so important, and finally discuss some of the limitations of RCTs with specific reference to external validity and generalizability.