RCTs are widespread in the medical arena. Here, the environment can be fully controlled.
Photo: WHO/Eduardo Soteras Jalil


<< First < Previous Page 1 Page 2 Page 3 Page 4 Page 5 Page 6 Page 7 Next > Last >>
Although randomised controlled trials are seeing widespread use, they have also been in for some criticism. Our author shows some of the snags that the method may meet with and recommends that context and appropriateness be given more consideration in designing evaluations.

Randomised controlled trials (RCTs) have recently grown in popularity. The basic idea is simple. In a randomised study, individuals are randomly assigned to so-called treatment and control groups, whereby both groups must be drawn from individuals whom the programme has yet to serve, so that the impact of an entire programme can be evaluated.

This random assignment to either treatment or control groups ensures that potential outcomes are not contaminated by self-selection into treatment. Self-selection refers to individuals selecting themselves into participating in particular programmes, e.g. they may self-select into microfinance programmes because they are particularly entrepreneurial or have certain risk attitudes and/or business skills. If randomisation is successful, it is assumed that individuals in treatment and control groups are equivalent in terms of observable and unobservable characteristics, with the exception of the treatment status. As a result of this, the differences we observe in the outcomes of each of these individuals are understood to be the effect of the programme.

The crucial aspect of causality

Hype surrounding RCTs has led policy-makers, funders and researchers to believe that randomisation is the only method that convincingly establishes causality.

<< First < Previous Page 1 Page 2 Page 3 Page 4 Page 5 Page 6 Page 7 Next > Last >>