Treatment noncompliance in randomized experiments

Statistical approaches and design issues

Brad J. Sagarin, Stephen West, Alexander Ratnikov, William K. Homan, Timothy D. Ritchie, Edward J. Hansen

Research output: Contribution to journalArticle

14 Citations (Scopus)

Abstract

Treatment noncompliance in randomized experiments threatens the validity of causal inference and the interpretability of treatment effects. This article provides a nontechnical review of 7 approaches: 3 traditional and 4 newer statistical analysis strategies. Traditional approaches include (a) intention-to-treat analysis (which estimates the effects of treatment assignment irrespective of treatment received), (b) as-treated analysis (which reassigns participants to groups reflecting the treatment they actually received), and (c) per-protocol analysis (which drops participants who did not comply with their assigned treatment). Newer approaches include (d) the complier average causal effect (which estimates the effect of treatment on the subpopulation of those who would comply with their assigned treatment), (e) dose-response estimation (which uses degree of compliance to stratify participants, producing an estimate of a dose-response relationship), (f) propensity score analysis (which uses covariates to estimate the probability that individual participants will comply, enabling estimates of treatment effects at different propensities), and (g) treatment effect bounding (which calculates a range of possible treatment effects applicable to both compliers and noncompliers). The discussion considers the areas of application, the quantity estimated, the underlying assumptions, and the strengths and weaknesses of each approach.

Original languageEnglish (US)
Pages (from-to)317-333
Number of pages17
JournalPsychological Methods
Volume19
Issue number3
DOIs
StatePublished - 2014

Fingerprint

Propensity Score
Intention to Treat Analysis
Compliance
Randomized Experiments
Treatment Effects

Keywords

  • Causal inference
  • Intention to treat
  • Nonadherence
  • Noncompliance
  • Potential outcomes

ASJC Scopus subject areas

  • Psychology (miscellaneous)
  • History and Philosophy of Science

Cite this

Sagarin, B. J., West, S., Ratnikov, A., Homan, W. K., Ritchie, T. D., & Hansen, E. J. (2014). Treatment noncompliance in randomized experiments: Statistical approaches and design issues. Psychological Methods, 19(3), 317-333. https://doi.org/10.1037/met0000013

Treatment noncompliance in randomized experiments : Statistical approaches and design issues. / Sagarin, Brad J.; West, Stephen; Ratnikov, Alexander; Homan, William K.; Ritchie, Timothy D.; Hansen, Edward J.

In: Psychological Methods, Vol. 19, No. 3, 2014, p. 317-333.

Research output: Contribution to journalArticle

Sagarin, Brad J. ; West, Stephen ; Ratnikov, Alexander ; Homan, William K. ; Ritchie, Timothy D. ; Hansen, Edward J. / Treatment noncompliance in randomized experiments : Statistical approaches and design issues. In: Psychological Methods. 2014 ; Vol. 19, No. 3. pp. 317-333.
@article{dd4ae177b0a64a2f995ef992e8480398,
title = "Treatment noncompliance in randomized experiments: Statistical approaches and design issues",
abstract = "Treatment noncompliance in randomized experiments threatens the validity of causal inference and the interpretability of treatment effects. This article provides a nontechnical review of 7 approaches: 3 traditional and 4 newer statistical analysis strategies. Traditional approaches include (a) intention-to-treat analysis (which estimates the effects of treatment assignment irrespective of treatment received), (b) as-treated analysis (which reassigns participants to groups reflecting the treatment they actually received), and (c) per-protocol analysis (which drops participants who did not comply with their assigned treatment). Newer approaches include (d) the complier average causal effect (which estimates the effect of treatment on the subpopulation of those who would comply with their assigned treatment), (e) dose-response estimation (which uses degree of compliance to stratify participants, producing an estimate of a dose-response relationship), (f) propensity score analysis (which uses covariates to estimate the probability that individual participants will comply, enabling estimates of treatment effects at different propensities), and (g) treatment effect bounding (which calculates a range of possible treatment effects applicable to both compliers and noncompliers). The discussion considers the areas of application, the quantity estimated, the underlying assumptions, and the strengths and weaknesses of each approach.",
keywords = "Causal inference, Intention to treat, Nonadherence, Noncompliance, Potential outcomes",
author = "Sagarin, {Brad J.} and Stephen West and Alexander Ratnikov and Homan, {William K.} and Ritchie, {Timothy D.} and Hansen, {Edward J.}",
year = "2014",
doi = "10.1037/met0000013",
language = "English (US)",
volume = "19",
pages = "317--333",
journal = "Psychological Methods",
issn = "1082-989X",
publisher = "American Psychological Association Inc.",
number = "3",

}

TY - JOUR

T1 - Treatment noncompliance in randomized experiments

T2 - Statistical approaches and design issues

AU - Sagarin, Brad J.

AU - West, Stephen

AU - Ratnikov, Alexander

AU - Homan, William K.

AU - Ritchie, Timothy D.

AU - Hansen, Edward J.

PY - 2014

Y1 - 2014

N2 - Treatment noncompliance in randomized experiments threatens the validity of causal inference and the interpretability of treatment effects. This article provides a nontechnical review of 7 approaches: 3 traditional and 4 newer statistical analysis strategies. Traditional approaches include (a) intention-to-treat analysis (which estimates the effects of treatment assignment irrespective of treatment received), (b) as-treated analysis (which reassigns participants to groups reflecting the treatment they actually received), and (c) per-protocol analysis (which drops participants who did not comply with their assigned treatment). Newer approaches include (d) the complier average causal effect (which estimates the effect of treatment on the subpopulation of those who would comply with their assigned treatment), (e) dose-response estimation (which uses degree of compliance to stratify participants, producing an estimate of a dose-response relationship), (f) propensity score analysis (which uses covariates to estimate the probability that individual participants will comply, enabling estimates of treatment effects at different propensities), and (g) treatment effect bounding (which calculates a range of possible treatment effects applicable to both compliers and noncompliers). The discussion considers the areas of application, the quantity estimated, the underlying assumptions, and the strengths and weaknesses of each approach.

AB - Treatment noncompliance in randomized experiments threatens the validity of causal inference and the interpretability of treatment effects. This article provides a nontechnical review of 7 approaches: 3 traditional and 4 newer statistical analysis strategies. Traditional approaches include (a) intention-to-treat analysis (which estimates the effects of treatment assignment irrespective of treatment received), (b) as-treated analysis (which reassigns participants to groups reflecting the treatment they actually received), and (c) per-protocol analysis (which drops participants who did not comply with their assigned treatment). Newer approaches include (d) the complier average causal effect (which estimates the effect of treatment on the subpopulation of those who would comply with their assigned treatment), (e) dose-response estimation (which uses degree of compliance to stratify participants, producing an estimate of a dose-response relationship), (f) propensity score analysis (which uses covariates to estimate the probability that individual participants will comply, enabling estimates of treatment effects at different propensities), and (g) treatment effect bounding (which calculates a range of possible treatment effects applicable to both compliers and noncompliers). The discussion considers the areas of application, the quantity estimated, the underlying assumptions, and the strengths and weaknesses of each approach.

KW - Causal inference

KW - Intention to treat

KW - Nonadherence

KW - Noncompliance

KW - Potential outcomes

UR - http://www.scopus.com/inward/record.url?scp=84925767664&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84925767664&partnerID=8YFLogxK

U2 - 10.1037/met0000013

DO - 10.1037/met0000013

M3 - Article

VL - 19

SP - 317

EP - 333

JO - Psychological Methods

JF - Psychological Methods

SN - 1082-989X

IS - 3

ER -