A control group is an untreated research sample against which all other groups or samples in the research is compared.
A control group is constructed to produce an estimate of the counter-factual, that is, what would have happened if an intervention had not been implemented. A control group is constructed by randomly assigning people to either the control group or to one or more "treatment" groups.
Sometimes the label "control group" is used to refer to any sort of comparison group, even when random assignment has not been used to create it. This is not recommended, as it confuses an important difference.
If random assignment is used to create the control group, it is likely that on average the characteristics of the control group will be the same as those of the treatment group - but it is possible that they will be different. It is therefore important to check their similarity on observable variables that might influence the outcomes and impacts (for example, important demographic variables such as gender, age, level of income).
This overview from the American Evaluation Association's AEA 365 blog page, provides an overview of using waitlist control groups. It includes a tip on how to use wait lists in a program that is going to use a pre and post-test.
This video guide, produced by UNICEF, summarises the key features of RCTs with a particular emphasis on their use in impact evaluation.
This guide, written by Howard White, Shagun Sabarwal and Thomas de Hoop for UNICEF, looks at the use of Randomized Control Trials (RCTs) in Impact Evaluation.
Randomised controlled trials (RCTs), or randomised impact evaluations, are a type of impact evaluation that uses randomised access to social programmes as a means of limiting bias and generating an internally valid impact estimate.