Regression discontinuity design

Regression discontinuity design (RDD) is a quasi-experimental research method used to identify causal effects by exploiting a cutoff or threshold in the assignment of treatment.
Written by
Reviewed by
Updated on Jun 12, 2024
Reading time 4 minutes

3 key takeaways:

Copy link to section
  • Regression discontinuity design leverages a known cutoff or threshold to assign treatment, creating a natural experiment that helps identify causal effects.
  • RDD provides robust causal inference by comparing outcomes for individuals just above and just below the threshold, assuming they are similar except for the treatment.
  • It is widely used to assess the impact of interventions in policy evaluation, education, health economics, and other fields.

What is regression discontinuity design?

Copy link to section

Regression discontinuity design (RDD) is a statistical technique used to estimate causal effects by assigning a treatment based on whether an observed variable crosses a predetermined threshold.

This method assumes that individuals near the threshold on either side are similar in all respects except for the treatment assignment, allowing for a comparison that can reveal the causal effect of the treatment.

For example, in an educational context, a scholarship program might be awarded to students who score above a certain test score threshold. By comparing students who scored just above and just below the cutoff, researchers can infer the causal impact of receiving the scholarship on educational outcomes.

Key components of regression discontinuity design

Copy link to section

Several key components define the regression discontinuity design:

  • Threshold (Cutoff): The point at which the assignment of treatment changes. This threshold is crucial for creating the discontinuity used for causal inference.
  • Running Variable: The continuous variable used to determine treatment assignment, such as a test score or income level.
  • Outcome Variable: The dependent variable that researchers are interested in studying to understand the causal impact of the treatment.

By focusing on the discontinuity at the threshold, RDD isolates the local causal effect of the treatment.

Applications of regression discontinuity design

Copy link to section

Regression discontinuity design is widely applied in various fields to study causal relationships:

  • Education: Evaluating the impact of remedial programs, scholarships, or admission policies that have clear cutoff points.
  • Public Policy: Assessing the effects of welfare programs, tax policies, or regulatory changes implemented based on specific eligibility criteria.
  • Health Economics: Investigating the impact of health interventions, such as access to insurance or medical treatments, where eligibility is determined by an age or income threshold.

These applications demonstrate the versatility of RDD in providing causal insights across different policy contexts.

Advantages of regression discontinuity design

Copy link to section

Regression discontinuity design offers several advantages:

  • Causal Inference: RDD provides strong causal evidence by leveraging the randomness around the cutoff, making the treatment and control groups comparable.
  • Clarity: The clear assignment rule based on the threshold makes the identification strategy straightforward and transparent.
  • Robustness: RDD is less susceptible to certain biases that can affect other observational study designs, as long as the assumptions hold near the threshold.

These advantages make RDD a valuable tool for empirical research and policy evaluation.

Challenges of regression discontinuity design

Copy link to section

Despite its strengths, regression discontinuity design also faces several challenges:

  • Data Requirements: RDD requires a large sample size around the cutoff to provide precise estimates of the treatment effect.
  • Assumptions: The method relies on the assumption that individuals just above and below the threshold are similar except for the treatment. Any manipulation of the running variable near the threshold can bias results.
  • Local Effect: RDD identifies the local average treatment effect around the threshold, which may not generalize to the entire population.

Addressing these challenges is essential for the valid application of RDD in empirical research.

Examples of regression discontinuity design in action

Copy link to section

Here are some examples illustrating the use of regression discontinuity design:

  • School Funding: Analyzing the impact of additional school funding allocated to schools with enrollment numbers just above a cutoff.
  • Minimum Wage Laws: Investigating the effects of minimum wage increases in regions where the policy is implemented based on specific economic indicators crossing a threshold.
  • Healthcare Access: Studying the impact of Medicaid expansion on health outcomes by comparing individuals just above and below the income eligibility cutoff.

These examples highlight how RDD can be applied to derive causal inferences from policy-induced changes in various fields.

Exploring related concepts such as instrumental variables, difference-in-differences, causal inference, and natural experiments can provide further insights into the principles and applications of regression discontinuity design in empirical research.


Sources & references

Arti

Arti

AI Financial Assistant

  • Finance
  • Investing
  • Trading
  • Stock Market
  • Cryptocurrency
Arti is a specialized AI Financial Assistant at Invezz, created to support the editorial team. He leverages both AI and the Invezz.com knowledge base, understands over 100,000 Invezz related data points, has read every piece of research, news and guidance we\'ve ever produced, and is trained to never make up new...