A literature review of the use of random assignment methodology in evaluations of US social policy programmes

The (then) Department of Social Security commissioned CRSP to undertake a brief review of the use of social experiments in evaluations of social security, welfare-to-work, education and training and other relevant social policies. The review focuses on potential difficulties with implementing and operating random assignment and the strategies and options for overcoming them. It was commissioned in the context of the extension for New Deal for Disabled People. Social experiments provide the estimate of the impact of a programme, the difference between what happens and what would have happened in the absence of the programme. They involve the random assignment of individuals to at least one treatment group and a control group. The advantages and disadvantages of social experiments (Section 1.1) and their uses are summarised (Section 1.1.1).