Randomized Control Trial

WHY RCTs ARE USEFUL

Assessing the effectiveness of different types of interventions is a major challenge for researchers, scientists, policymakers, and the public generally. Does intensive social work input for children with behavior problems actually help these children and their families? What is the impact on children’s literacy of computer-assisted learning? Does introducing criminals to their victims secure lower reoffense rates? Do screening programs for breast or prostate cancer work in reducing deaths from these diseases? Answering such questions requires the use of a research design that permits the impact of the intervention in question to be disentangled, so far as is possible, from other influences, including the social characteristics of the intervention recipients and whatever else is going on in the social context at the time. RCTs can provide rigorous evidence to inform policy and practice decisions about the relative effectiveness of different interventions. As in any research, a good design is only optimally useful if it addresses an important question and is well operationalized.

THE HISTORY OF RCTs

Efforts to control biases in assessing the effects of interventions go back at least three centuries, include both medical and social science, and are based on the important recognition that even well-intentioned interventions may do harm. Physicians from the 17th century on recommended casting lots to determine which patients would receive a particular treatment. This recognition of the need to compare like with like gathered momentum during the 18th century, such as naval surgeon James Lind’s famous study of scurvy, in which patients with the same condition were assigned to different treatments, and those who got oranges and limes recovered most quickly. The use of alternate or “random” allocation to generate comparison groups became more common after about 1920, with the first detailed descriptions of random allocation processes occurring in the 1930s. The term control was used in experimental psychology from the 1870s on to mean a standard of comparison used to check inferences made from an experiment; it comes from “counter-roll,” a duplicate register made to verify an official account.

Although today, RCTs are mainly associated with health care, social scientists have been equally concerned with designing intervention evaluation so as to be able to answer impact questions reliably. In the 1920s, 1930s, and 1940s, researchers, particularly in North America, in education, social welfare, and public policy made extensive use of experimental research designs, and “experimental sociology” was an important field of study. This was followed by “the golden age of evaluation,” from the 1960s to the 1980s, when American evaluators, often with government backing, used RCTs to evaluate a wide range of public policy initiatives, including income maintenance and housing allowance schemes, programs of support for disadvantaged workers, and interventions for former prison inmates.

...