• Entry
  • Reader's guide
  • Entries A-Z
  • Subject index

Regression-Discontinuity Design

A regression-discontinuity design (RDD) is an approach for estimating the impact of a program or the effect of a policy when a randomized experiment is not possible. When an RDD can be appropriately implemented, it is viewed as perhaps the most rigorous possible alternative to an experimental design for estimating the impact of an educational intervention. The key to the rigor of the RDD is that the selection mechanism—the process of determining who receives an educational intervention—is fully known and accounted for in the estimation process. This allows RDDs to avoid the problem to which most nonexperimental designs are highly susceptible—selection bias—whereby program impact estimates are confounded by unobserved differences between program participants and nonparticipants that are related to later outcomes. When feasible, RDDs ...

    • Loading...
    locked icon

    Sign in to access this content

    Get a 30 day FREE TRIAL

    • Watch videos from a variety of sources bringing classroom topics to life
    • Read modern, diverse business cases
    • Explore hundreds of books and reference titles