Regression Discontinuity with Covariates
Regression discontinuity with covariates estimates a causal effect at a cutoff of a running variable that determines treatment, adding pre-treatment covariates to improve precision without changing the identified effect.
공변량을 포함한 회귀불연속 설계는 처치를 결정하는 배정변수의 기준점에서 인과효과를 추정하되, 처치 이전 공변량을 추가하여 식별되는 효과는 바꾸지 않고 정밀도를 높인다.
How do I estimate a treatment effect when treatment is assigned by a score cutoff?
점수 기준선으로 처치가 배정될 때 처치효과를 어떻게 추정하나요?
Regression discontinuity estimates the effect of a treatment assigned by a cutoff on a running variable, such as a test score or age, by comparing units just above and below the threshold. Adding pre-treatment covariates can shrink standard errors, but identification still rests on continuity of potential outcomes at the cutoff.
When to use Regression Discontinuity with Covariates
- Treatment assigned by a known cutoff on a continuous running variable
- An outcome measured for units on both sides of the cutoff
- Pre-treatment covariates available to improve precision
When to use something else
- Treatment assignment depends on a cutoff only partially; use fuzzy regression discontinuity
- No assignment rule exists; use propensity score matching or difference-in-differences
Assumptions and requirements
- No precise manipulation of the running variable (density test)
- Continuity of potential outcomes at the cutoff
- Covariates determined before treatment and balanced at the cutoff
- Appropriate bandwidth and local polynomial order
What the analysis reports
- Local treatment effect at the cutoff with robust bias-corrected confidence interval
- Bandwidth and number of observations on each side
- Regression discontinuity plot
- Density manipulation test
- Covariate balance tests at the cutoff
How to report it
The covariate-adjusted regression discontinuity estimate was [value] (robust 95% CI [[value], [value]]) using a bandwidth of [value].
Related methods
- Regression Discontinuity Design
- Fuzzy Regression Discontinuity
- Difference-in-Differences
Run Regression Discontinuity with Covariates on Quorant
- Data you need
- a table (CSV or Excel) with an outcome variable, running and predictor (feature) columns
- Price
- 250 credits for one analysis up to 10,000 rows (complexity level 2); larger or wider data costs more. Optional methodologist review: 1200 credits.
- Checking
- Python results are cross-checked against R reference outputs in every deployment
- Before you pay
- A free pre-analysis diagnosis checks your data and gives an exact quote; nothing is deducted until you approve it.