A comparison of subset selection and analysis of covariance for the adjustment of confounders

Roderick J. Little, Hyonggin An, Jewel Johanns, Bruno Giordani

Research output: Contribution to journalReview article

15 Citations (Scopus)

Abstract

Two common methods for adjusting group comparisons for differences in the distribution of confounders, namely analysis of covariance (ANCOVA) and subset selection, are compared using real examples from neuropsychology, theory, and simulations. ANCOVA has potential pitfalls, but the blanket rejection of the method in some areas of empirical psychology is not justified. Assumptions of the methods are reviewed, with issues of selection bias, nonlinearity, and interaction emphasized. Advantages of ANCOVA include better power, improved ability to detect and estimate interactions, and the availability of extensions to deal with measurement error in the covariates. Forms of ANCOVA are advocated that relax the standard assumption of linearity between the outcome and covariates. Specifically, a version of ANCOVA that models the relationship between the covariate and the outcome through cubic spline with fixed knots outperforms other methods in simulations.

Original languageEnglish
Pages (from-to)459-476
Number of pages18
JournalPsychological Methods
Volume5
Issue number4
DOIs
Publication statusPublished - 2000 Jan 1
Externally publishedYes

Fingerprint

Social Adjustment
Neuropsychology
Aptitude
Selection Bias
Psychology

ASJC Scopus subject areas

  • Psychology (miscellaneous)

Cite this

A comparison of subset selection and analysis of covariance for the adjustment of confounders. / Little, Roderick J.; An, Hyonggin; Johanns, Jewel; Giordani, Bruno.

In: Psychological Methods, Vol. 5, No. 4, 01.01.2000, p. 459-476.

Research output: Contribution to journalReview article

Little, Roderick J. ; An, Hyonggin ; Johanns, Jewel ; Giordani, Bruno. / A comparison of subset selection and analysis of covariance for the adjustment of confounders. In: Psychological Methods. 2000 ; Vol. 5, No. 4. pp. 459-476.
@article{5c50864b6bc84146bab682ad93bd92ab,
title = "A comparison of subset selection and analysis of covariance for the adjustment of confounders",
abstract = "Two common methods for adjusting group comparisons for differences in the distribution of confounders, namely analysis of covariance (ANCOVA) and subset selection, are compared using real examples from neuropsychology, theory, and simulations. ANCOVA has potential pitfalls, but the blanket rejection of the method in some areas of empirical psychology is not justified. Assumptions of the methods are reviewed, with issues of selection bias, nonlinearity, and interaction emphasized. Advantages of ANCOVA include better power, improved ability to detect and estimate interactions, and the availability of extensions to deal with measurement error in the covariates. Forms of ANCOVA are advocated that relax the standard assumption of linearity between the outcome and covariates. Specifically, a version of ANCOVA that models the relationship between the covariate and the outcome through cubic spline with fixed knots outperforms other methods in simulations.",
author = "Little, {Roderick J.} and Hyonggin An and Jewel Johanns and Bruno Giordani",
year = "2000",
month = "1",
day = "1",
doi = "10.1037/1082-989X.5.4.459",
language = "English",
volume = "5",
pages = "459--476",
journal = "Psychological Methods",
issn = "1082-989X",
publisher = "American Psychological Association Inc.",
number = "4",

}

TY - JOUR

T1 - A comparison of subset selection and analysis of covariance for the adjustment of confounders

AU - Little, Roderick J.

AU - An, Hyonggin

AU - Johanns, Jewel

AU - Giordani, Bruno

PY - 2000/1/1

Y1 - 2000/1/1

N2 - Two common methods for adjusting group comparisons for differences in the distribution of confounders, namely analysis of covariance (ANCOVA) and subset selection, are compared using real examples from neuropsychology, theory, and simulations. ANCOVA has potential pitfalls, but the blanket rejection of the method in some areas of empirical psychology is not justified. Assumptions of the methods are reviewed, with issues of selection bias, nonlinearity, and interaction emphasized. Advantages of ANCOVA include better power, improved ability to detect and estimate interactions, and the availability of extensions to deal with measurement error in the covariates. Forms of ANCOVA are advocated that relax the standard assumption of linearity between the outcome and covariates. Specifically, a version of ANCOVA that models the relationship between the covariate and the outcome through cubic spline with fixed knots outperforms other methods in simulations.

AB - Two common methods for adjusting group comparisons for differences in the distribution of confounders, namely analysis of covariance (ANCOVA) and subset selection, are compared using real examples from neuropsychology, theory, and simulations. ANCOVA has potential pitfalls, but the blanket rejection of the method in some areas of empirical psychology is not justified. Assumptions of the methods are reviewed, with issues of selection bias, nonlinearity, and interaction emphasized. Advantages of ANCOVA include better power, improved ability to detect and estimate interactions, and the availability of extensions to deal with measurement error in the covariates. Forms of ANCOVA are advocated that relax the standard assumption of linearity between the outcome and covariates. Specifically, a version of ANCOVA that models the relationship between the covariate and the outcome through cubic spline with fixed knots outperforms other methods in simulations.

UR - http://www.scopus.com/inward/record.url?scp=0034576526&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0034576526&partnerID=8YFLogxK

U2 - 10.1037/1082-989X.5.4.459

DO - 10.1037/1082-989X.5.4.459

M3 - Review article

C2 - 11194208

AN - SCOPUS:0034576526

VL - 5

SP - 459

EP - 476

JO - Psychological Methods

JF - Psychological Methods

SN - 1082-989X

IS - 4

ER -