The Simpson's paradox unraveled

Miguel A Hernán; David Clayton; Niels Keiding

doi:10.1093/ije/dyr041

The Simpson's paradox unraveled

Miguel A Hernán, David Clayton, Niels Keiding

90 Citations (Scopus)

Abstract

Background: In a famous article, Simpson described a hypothetical data example that led to apparently paradoxical results. Methods: We make the causal structure of Simpson's example explicit. Results: We show how the paradox disappears when the statistical analysis is appropriately guided by subject-matter knowledge. We also review previous explanations of Simpson's paradox that attributed it to two distinct phenomena: confounding and non-collapsibility. Conclusion: Analytical errors may occur when the problem is stripped of its causal context and analyzed merely in statistical terms. Published by Oxford University Press on behalf of the International Epidemiological Association.

Original language	Danish
Journal	International Journal of Epidemiology
Volume	40
Issue number	3
Pages (from-to)	780-785
Number of pages	6
ISSN	0300-5771
DOIs	https://doi.org/10.1093/ije/dyr041
Publication status	Published - Jun 2011

Access to Document

10.1093/ije/dyr041

Cite this

@article{81f720626196486a956310410bbbbaeb,

title = "The Simpson's paradox unraveled",

abstract = "Background: In a famous article, Simpson described a hypothetical data example that led to apparently paradoxical results. Methods: We make the causal structure of Simpson's example explicit. Results: We show how the paradox disappears when the statistical analysis is appropriately guided by subject-matter knowledge. We also review previous explanations of Simpson's paradox that attributed it to two distinct phenomena: confounding and non-collapsibility. Conclusion: Analytical errors may occur when the problem is stripped of its causal context and analyzed merely in statistical terms. Published by Oxford University Press on behalf of the International Epidemiological Association.",

author = "Hern{\'a}n, {Miguel A} and David Clayton and Niels Keiding",

year = "2011",

month = jun,

doi = "10.1093/ije/dyr041",

language = "Dansk",

volume = "40",

pages = "780--785",

journal = "International Journal of Epidemiology",

issn = "0300-5771",

publisher = "Oxford University Press",

number = "3",

}

TY - JOUR

T1 - The Simpson's paradox unraveled

AU - Hernán, Miguel A

AU - Clayton, David

AU - Keiding, Niels

PY - 2011/6

Y1 - 2011/6

N2 - Background: In a famous article, Simpson described a hypothetical data example that led to apparently paradoxical results. Methods: We make the causal structure of Simpson's example explicit. Results: We show how the paradox disappears when the statistical analysis is appropriately guided by subject-matter knowledge. We also review previous explanations of Simpson's paradox that attributed it to two distinct phenomena: confounding and non-collapsibility. Conclusion: Analytical errors may occur when the problem is stripped of its causal context and analyzed merely in statistical terms. Published by Oxford University Press on behalf of the International Epidemiological Association.

AB - Background: In a famous article, Simpson described a hypothetical data example that led to apparently paradoxical results. Methods: We make the causal structure of Simpson's example explicit. Results: We show how the paradox disappears when the statistical analysis is appropriately guided by subject-matter knowledge. We also review previous explanations of Simpson's paradox that attributed it to two distinct phenomena: confounding and non-collapsibility. Conclusion: Analytical errors may occur when the problem is stripped of its causal context and analyzed merely in statistical terms. Published by Oxford University Press on behalf of the International Epidemiological Association.

U2 - 10.1093/ije/dyr041

DO - 10.1093/ije/dyr041

M3 - Tidsskriftartikel

SN - 0300-5771

VL - 40

SP - 780

EP - 785

JO - International Journal of Epidemiology

JF - International Journal of Epidemiology

IS - 3

ER -