“Incorporating Self-Selection into the Random Assignment Experiment” -- IGPA professors James Kuklinski (left) and Brian Gaines (right). 

Political scientists sometimes use the random assignment experiment to make causal inferences about the real world. An advantage of randomization is that, in principle, it eliminates the effects of all factors other than the treatment. Self-selection is ubiquitous in the real-world, however, and for some purposes including a self-selection condition in the experiment can prove helpful, providing insights that the random assignment experiment alone cannot.