Noise-Induced Randomization in Regression Discontinuity Designs

Regression discontinuity designs are used to estimate causal effects in settings where treatment is determined by whether an observed running variable crosses a pre-specified threshold. While the resulting sampling design is sometimes described as akin to a locally randomized experiment in a neighborhood of the threshold, standard formal analyses do not make reference to probabilistic treatment assignment and instead identify treatment effects via continuity arguments. Here we propose a new approach to identification, estimation, and inference in regression discontinuity designs that exploits measurement error in the running variable. Under an assumption that the measurement error is exogenous, we show how to consistently estimate causal effects using a class of linear estimators that weight treated and control units so as to balance a latent variable of which the running variable is a noisy measure. We find this approach to facilitate identification of both familiar estimands from the literature, as well as policy-relevant estimands that correspond to the effects of realistic changes to the existing treatment assignment rule. We demonstrate the method with a study of retention of HIV patients and evaluate its performance using simulated data and a regression discontinuity design artificially constructed from test scores in early childhood.

[1]  Alessandra Mattei,et al.  Regression Discontinuity Designs as Local Randomized Experiments , 2021 .

[2]  Luke W. Miratrix,et al.  Do School Districts Affect NYC House Prices? Identifying Border Differences Using a Bayesian Nonparametric Approach to Geographic Regression Discontinuity Designs , 2020 .

[3]  Mhairi Maskew,et al.  Impact of Viral Load Monitoring on Retention and Viral Suppression: A Regression Discontinuity Analysis of South Africa's National Laboratory Cohort. , 2020, American journal of epidemiology.

[4]  Juan D. Diaz,et al.  Complex Discontinuity Designs Using Covariates , 2020 .

[5]  Otávio Bartalotti,et al.  A Correction for Regression Discontinuity Designs With Group-Specific Mismeasurement of the Running Variable , 2020, Journal of Business & Economic Statistics.

[6]  Ji Seung Yang,et al.  Latent Variable Regression Discontinuity Design with Cluster Level Treatment Assignment , 2019, Multivariate behavioral research.

[7]  Morgan Kelly Persistence , 2019, Encyclopedia of Evolutionary Psychological Science.

[8]  Zhichao Jiang,et al.  Measurement errors in the binary instrumental variable model , 2019, Biometrika.

[9]  Fan Li,et al.  A regression discontinuity design for ordinal running variables: Evaluating central bank purchases of corporate bonds , 2019, The Annals of Applied Statistics.

[10]  Stefan Wager,et al.  Confidence Intervals for Nonparametric Empirical Bayes Analysis , 2019, Journal of the American Statistical Association.

[11]  Stefan Wager,et al.  Bias-Aware Confidence Intervals for Empirical Bayes Analysis , 2019 .

[12]  Hal Varian,et al.  Optimizing the tie-breaker regression discontinuity design , 2018, Electronic Journal of Statistics.

[13]  Luke Miratrix,et al.  A Bayesian Nonparametric Approach to Geographic Regression Discontinuity Designs: Do School Districts Affect NYC House Prices? , 2018, 1807.04516.

[14]  Deborah K Glencross,et al.  CD4 cell count variability with repeat testing in South Africa: Should reporting include both absolute counts and ranges of plausible values? , 2018, International journal of STD & AIDS.

[15]  Simon Jäger,et al.  A Permutation Test for the Regression Kink Design , 2018 .

[16]  Till Bärnighausen,et al.  Treatment eligibility and retention in clinical HIV care: A regression discontinuity study in South Africa , 2017, PLoS medicine.

[17]  A. Cooperman Randomization Inference with Rainfall Data: Using Historical Weather Patterns for Variance Estimation , 2017, Political Analysis.

[18]  Magne Mogstad,et al.  Beyond LATE with a Discrete Instrument , 2017, Journal of Political Economy.

[19]  Jasjeet S. Sekhon,et al.  On Interpreting the Regression Discontinuity Design as a Local Experiment , 2017 .

[20]  Stefan Wager,et al.  Optimized Regression Discontinuity Designs , 2017, Review of Economics and Statistics.

[21]  Luke W. Miratrix,et al.  A nonparametric Bayesian methodology for regression discontinuity designs , 2017, Journal of Statistical Planning and Inference.

[22]  Thomas Le Barbanchon,et al.  Regression Discontinuity Design with Continuous Measurement Error in the Running Variable , 2017, SSRN Electronic Journal.

[23]  Nathan Kallus,et al.  Generalized Optimal Matching Methods for Causal Inference , 2016, J. Mach. Learn. Res..

[24]  Lucas Janson,et al.  Panning for gold: ‘model‐X’ knockoffs for high dimensional controlled variable selection , 2016, 1610.02351.

[25]  Zhuan Pei,et al.  The Devil is in the Tails: Regression Discontinuity Design with Measurement Error in the Assignment Variable , 2016, SSRN Electronic Journal.

[26]  Christoph Rothe,et al.  Inference in Regression Discontinuity Designs with a Discrete Running Variable , 2016, American Economic Review.

[27]  Timothy B. Armstrong,et al.  Simple and Honest Confidence Intervals in Nonparametric Regression , 2016, Quantitative Economics.

[28]  Bernard L. Fraga,et al.  Examining the Causal Impact of the Voting Rights Act Language Minority Provisions , 2016 .

[29]  Arthur Lewbel,et al.  Identifying the Effect of Changing the Policy Threshold in Regression Discontinuity Models , 2015, Review of Economics and Statistics.

[30]  Timothy B. Armstrong,et al.  Optimal Inference in a Class of Regression Models , 2015, 1511.06028.

[31]  Joshua D. Angrist,et al.  Wanna Get Away? Regression Discontinuity Estimation of Exam School Effects Away From the Cutoff , 2015 .

[32]  Max H. Farrell,et al.  On the Effect of Bias Estimation on Coverage Accuracy in Nonparametric Inference , 2015, Journal of the American Statistical Association.

[33]  Iain Dunning,et al.  JuMP: A Modeling Language for Mathematical Optimization , 2015, SIAM Rev..

[34]  Alessandra Mattei,et al.  Evaluating the Causal Effect of University Grants on Student Dropout: Evidence from a Regression Discontinuity Design Using Principal Stratification , 2015, 1507.04199.

[35]  Michelle Najarian,et al.  Early Childhood Longitudinal Study, Kindergarten Class of 2010-11 (ECLS-K:2011). User's Manual for the ECLS-K:2011 Kindergarten Data File and Electronic Codebook, Public Version. NCES 2015-074. , 2015 .

[36]  Matias D. Cattaneo,et al.  Randomization Inference in the Regression Discontinuity Design: An Application to Party Advantages in the U.S. Senate , 2015 .

[37]  Takahide Yanagi,et al.  The Effect of Measurement Error in the Sharp Regression Discontinuity Design , 2014 .

[38]  Alan Edelman,et al.  Julia: A Fresh Approach to Numerical Computing , 2014, SIAM Rev..

[39]  Sebastian Calonico,et al.  Robust Nonparametric Confidence Intervals for Regression‐Discontinuity Designs , 2014 .

[40]  Aaron Roth,et al.  The Algorithmic Foundations of Differential Privacy , 2014, Found. Trends Theor. Comput. Sci..

[41]  Guido Imbens,et al.  Why High-Order Polynomials Should Not Be Used in Regression Discontinuity Designs , 2014 .

[42]  Till Bärnighausen,et al.  Regression Discontinuity Designs in Epidemiology , 2014, Epidemiology.

[43]  Luc Kestens,et al.  WHO Multicenter Evaluation of FACSCount CD4 and Pima CD4 T-Cell Count Systems: Instrument Performance and Misclassification of HIV-Infected Patients , 2014, Journal of acquired immune deficiency syndromes.

[44]  Kari Lock Morgan,et al.  Balancing Covariates via Propensity Score Weighting , 2014, 1609.07494.

[45]  Ben B. Hansen,et al.  Limitless Regression Discontinuity , 2014, Journal of Educational and Behavioral Statistics.

[46]  Gianluca Baio,et al.  Bayesian regression discontinuity designs: incorporating clinical knowledge in the causal analysis of primary care data , 2014, Statistics in medicine.

[47]  Arlene K. H. Kim Minimax bounds for estimation of normal mixtures , 2011, 1112.4565.

[48]  Jasjeet S. Sekhon,et al.  Elections and the Regression Discontinuity Design: Lessons from Close U.S. House Races, 1942–2008 , 2011, Political Analysis.

[49]  Judea Pearl,et al.  On Measurement Bias in Causal Inference , 2010, UAI.

[50]  A. Banerjee,et al.  Targeting the Poor: Evidence from a Field Experiment in Indonesia , 2010, The American economic review.

[51]  Wenhua Jiang,et al.  General maximum likelihood empirical Bayes estimation of normal means , 2009, 0908.1709.

[52]  A. Meister Deconvolution Problems in Nonparametric Statistics , 2009 .

[53]  Richard K. Crump,et al.  Dealing with limited overlap in estimation of average treatment effects , 2009 .

[54]  David Lee,et al.  Regression Discontinuity Designs in Economics , 2009 .

[55]  D. Rubin For objective causal inference, design trumps analysis , 2008, 0811.1640.

[56]  Thomas D. Cook,et al.  "Waiting for Life to Arrive": A history of the regression-discontinuity design in Psychology, Statistics and Economics , 2008 .

[57]  Harold P. Benson,et al.  A simplicial branch and bound duality-bounds algorithm for the linear sum-of-ratios problem , 2007, Eur. J. Oper. Res..

[58]  George A. Krause,et al.  The Republicans Should Pray for Rain: Weather, Turnout, and Voting in U.S. Presidential Elections , 2007, The Journal of Politics.

[59]  David S. Lee Randomized experiments from non-random selection in U.S. House elections , 2005 .

[60]  Charles F. Manski,et al.  Confidence Intervals for Partially Identified Parameters , 2003 .

[61]  D. Rubin,et al.  Principal Stratification in Causal Inference , 2002, Biometrics.

[62]  Hiroshi Konno,et al.  Minimization of the sum of three linear fractional functions , 1999, J. Glob. Optim..

[63]  H Robbins,et al.  Clinical and prophylactic trials with assured new treatment for those at greater risk: II. Examples. , 1996, American journal of public health.

[64]  H. Robbins,et al.  Clinical and prophylactic trials with assured new treatment for those at greater risk: I. A design proposal. , 1996, American journal of public health.

[65]  D. Donoho Statistical Estimation and Optimal Recovery , 1994 .

[66]  Joseph C. Cappelleri,et al.  Random Measurement Error Does Not Bias the Treatment Effect Estimate in the Regression-Discontinuity Design , 1991 .

[67]  H. Robbins,et al.  Estimating a multiplicative treatment effect under biased allocation , 1991 .

[68]  P. Massart The Tight Constant in the Dvoretzky-Kiefer-Wolfowitz Inequality , 1990 .

[69]  H. Robbins,et al.  Estimating the superiority of a drug to a placebo when all and only those patients at risk are treated with the drug. , 1989, Proceedings of the National Academy of Sciences of the United States of America.

[70]  H. Robbins,et al.  Estimating a treatment effect under biased sampling. , 1988, Proceedings of the National Academy of Sciences of the United States of America.

[71]  D. Rubin,et al.  The central role of the propensity score in observational studies for causal effects , 1983 .

[72]  D. Rubin Estimating causal effects of treatments in randomized and nonrandomized studies. , 1974 .

[73]  A. Charnes,et al.  Programming with linear fractional functionals , 1962 .

[74]  D. Campbell,et al.  Regression-Discontinuity Analysis: An Alternative to the Ex-Post Facto Experiment , 1960 .

[75]  J. Kiefer,et al.  CONSISTENCY OF THE MAXIMUM LIKELIHOOD ESTIMATOR IN THE PRESENCE OF INFINITELY MANY INCIDENTAL PARAMETERS , 1956 .

[76]  A. Roy Some thoughts on the distribution of earnings , 1951 .

[77]  Monica Morell Multilevel Regression Discontinuity Models with Latent Variables , 2020 .

[78]  R. Koenker,et al.  REBayes : Empirical Bayes Mixture Methods in R , 2017 .

[79]  Andrés Santos,et al.  USING INSTRUMENTAL VARIABLES FOR INFERENCE ABOUT POLICY RELEVANT TREATMENT PARAMETERS , 2016 .

[80]  L. Keele,et al.  Geographic Boundaries as Regression Discontinuities , 2015, Political Analysis.

[81]  Matias D. Cattaneo,et al.  rdrobust: An R Package for Robust Nonparametric Inference in Regression-Discontinuity Designs , 2015, R J..

[82]  Miikka A. T. Rokkanen,et al.  Exam Schools, Ability, and the Effects of Affirmative Action: Latent Factor Extrapolation in the Regression Discontinuity Design , 2015 .

[83]  J. Sekhon,et al.  Randomization Inference in the Regression Discontinuity Design: An Application to Party Advantages in the U.S. Senate ∗ , 2013 .

[84]  N. Léchopier " Experimental and quasi-experimental designs for research on teaching ", de Donald T. Campbell & Julian C. Stanley, (1963). , 2011 .

[85]  G. Janossy,et al.  Large‐scale affordable Panleucogated CD4+ testing with proactive internal and external quality assessment: In support of the South African national comprehensive care, treatment and management programme for HIV and AIDS , 2008, Cytometry. Part B, Clinical cytometry.

[86]  Petra E. Todd,et al.  IDENTIFICATION AND ESTIMATION OF TREATMENT EFFECTS WITH A REGRESSION-DISCONTINUITY DESIGN , 2001 .

[87]  M. Bilodeau Estimating a multivariate treatment effect under a biased allocation rule , 1997 .

[88]  Jianqing Fan,et al.  On automatic boundary corrections , 1997 .

[89]  D S Stein,et al.  Within-subject variation in CD4 lymphocyte count in asymptomatic human immunodeficiency virus infection: implications for patient monitoring. , 1994, The Journal of infectious diseases.

[90]  J. Heckman Sample selection bias as a specification error , 1979 .

[91]  H. Robbins An Empirical Bayes Approach to Statistics , 1956 .