Automated local regression discontinuity design discovery

William Herlands, Andrew Gordon Wilson, Edward McFowland, Daniel Neill

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Inferring causal relationships in observational data is crucial for understanding scientific and social processes. We develop the first statistical machine learning approach for automatically discovering regression discontinuity designs (RDDs), a quasi-experimental setup often used in econometrics. Our method identifies interpretable, localized RDDs in arbitrary dimensional data and can seamlessly compute treatment effects without expert supervision. By applying the technique to a variety of synthetic and real datasets, we demonstrate robust performance under adverse conditions including unobserved variables, substantial noise, and model misspecification.

Original languageEnglish (US)
Title of host publicationKDD 2018 - Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining
PublisherAssociation for Computing Machinery
Pages1512-1520
Number of pages9
ISBN (Print)9781450355520
DOIs
StatePublished - Jul 19 2018
Event24th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD 2018 - London, United Kingdom
Duration: Aug 19 2018Aug 23 2018

Other

Other24th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD 2018
CountryUnited Kingdom
CityLondon
Period8/19/188/23/18

Fingerprint

Learning systems

Keywords

  • Natural experiments
  • Pattern detection
  • Regression discontinuity

ASJC Scopus subject areas

  • Software
  • Information Systems

Cite this

Herlands, W., Wilson, A. G., McFowland, E., & Neill, D. (2018). Automated local regression discontinuity design discovery. In KDD 2018 - Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (pp. 1512-1520). Association for Computing Machinery. https://doi.org/10.1145/3219819.3219982

Automated local regression discontinuity design discovery. / Herlands, William; Wilson, Andrew Gordon; McFowland, Edward; Neill, Daniel.

KDD 2018 - Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Association for Computing Machinery, 2018. p. 1512-1520.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Herlands, W, Wilson, AG, McFowland, E & Neill, D 2018, Automated local regression discontinuity design discovery. in KDD 2018 - Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Association for Computing Machinery, pp. 1512-1520, 24th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD 2018, London, United Kingdom, 8/19/18. https://doi.org/10.1145/3219819.3219982
Herlands W, Wilson AG, McFowland E, Neill D. Automated local regression discontinuity design discovery. In KDD 2018 - Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Association for Computing Machinery. 2018. p. 1512-1520 https://doi.org/10.1145/3219819.3219982
Herlands, William ; Wilson, Andrew Gordon ; McFowland, Edward ; Neill, Daniel. / Automated local regression discontinuity design discovery. KDD 2018 - Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Association for Computing Machinery, 2018. pp. 1512-1520
@inproceedings{da39aaacf1ee49fca2911de42937b794,
title = "Automated local regression discontinuity design discovery",
abstract = "Inferring causal relationships in observational data is crucial for understanding scientific and social processes. We develop the first statistical machine learning approach for automatically discovering regression discontinuity designs (RDDs), a quasi-experimental setup often used in econometrics. Our method identifies interpretable, localized RDDs in arbitrary dimensional data and can seamlessly compute treatment effects without expert supervision. By applying the technique to a variety of synthetic and real datasets, we demonstrate robust performance under adverse conditions including unobserved variables, substantial noise, and model misspecification.",
keywords = "Natural experiments, Pattern detection, Regression discontinuity",
author = "William Herlands and Wilson, {Andrew Gordon} and Edward McFowland and Daniel Neill",
year = "2018",
month = "7",
day = "19",
doi = "10.1145/3219819.3219982",
language = "English (US)",
isbn = "9781450355520",
pages = "1512--1520",
booktitle = "KDD 2018 - Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining",
publisher = "Association for Computing Machinery",

}

TY - GEN

T1 - Automated local regression discontinuity design discovery

AU - Herlands, William

AU - Wilson, Andrew Gordon

AU - McFowland, Edward

AU - Neill, Daniel

PY - 2018/7/19

Y1 - 2018/7/19

N2 - Inferring causal relationships in observational data is crucial for understanding scientific and social processes. We develop the first statistical machine learning approach for automatically discovering regression discontinuity designs (RDDs), a quasi-experimental setup often used in econometrics. Our method identifies interpretable, localized RDDs in arbitrary dimensional data and can seamlessly compute treatment effects without expert supervision. By applying the technique to a variety of synthetic and real datasets, we demonstrate robust performance under adverse conditions including unobserved variables, substantial noise, and model misspecification.

AB - Inferring causal relationships in observational data is crucial for understanding scientific and social processes. We develop the first statistical machine learning approach for automatically discovering regression discontinuity designs (RDDs), a quasi-experimental setup often used in econometrics. Our method identifies interpretable, localized RDDs in arbitrary dimensional data and can seamlessly compute treatment effects without expert supervision. By applying the technique to a variety of synthetic and real datasets, we demonstrate robust performance under adverse conditions including unobserved variables, substantial noise, and model misspecification.

KW - Natural experiments

KW - Pattern detection

KW - Regression discontinuity

UR - http://www.scopus.com/inward/record.url?scp=85051562408&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85051562408&partnerID=8YFLogxK

U2 - 10.1145/3219819.3219982

DO - 10.1145/3219819.3219982

M3 - Conference contribution

SN - 9781450355520

SP - 1512

EP - 1520

BT - KDD 2018 - Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining

PB - Association for Computing Machinery

ER -