Inflation of the type I error rate when a continuous confounding variable is categorized in logistic regression analyses

被引:166
作者
Austin, PC
Brunner, LJ
机构
[1] Inst Clin Evaluat Sci, Toronto, ON M4N 3M5, Canada
[2] Univ Toronto, Dept Stat, Toronto, ON, Canada
[3] Univ Toronto, Dept Publ Hlth Sci, Toronto, ON, Canada
关键词
logistic regression; type I error rate; categorical variables; measurement error; confounding; epidemologic methods;
D O I
10.1002/sim.1687
中图分类号
Q [生物科学];
学科分类号
07 ; 0710 ; 09 ;
摘要
This paper demonstrates an inflation of the type I error rate that occurs when testing the statistical significance of a continuous risk factor after adjusting for a correlated continuous confounding variable that has been divided into a categorical variable. We used Monte Carlo simulation methods to assess the inflation of the type I error rate when testing the statistical significance of a risk factor after adjusting for a continuous confounding variable that has been divided into categories. We found that the inflation of the type I error rate increases with increasing sample size, as the correlation between the risk factor and the confounding variable increases, and with a decrease in the number of categories into which the confounder is divided. Even when the confounder is divided in a five-level categorical variable, the inflation of the type I error rate remained high when both the sample size and the correlation between the risk factor and the confounder were high. Copyright (C) 2004 John Wiley Sons, Ltd.
引用
收藏
页码:1159 / 1178
页数:20
相关论文
共 30 条
[1]   DANGERS OF USING OPTIMAL CUTPOINTS IN THE EVALUATION OF PROGNOSTIC FACTORS [J].
ALTMAN, DG ;
LAUSEN, B ;
SAUERBREI, W ;
SCHUMACHER, M .
JOURNAL OF THE NATIONAL CANCER INSTITUTE, 1994, 86 (11) :829-835
[2]  
Armitage P., 2002, STAT METHODS MED RES
[3]  
Arnold SF., 1990, Mathematical statistics
[4]   Type I error inflation in the presence of a ceiling effect [J].
Austin, PC ;
Brunner, LJ .
AMERICAN STATISTICIAN, 2003, 57 (02) :97-104
[5]   Estimating linear regression models in the presence of a censored independent variable [J].
Austin, PC ;
Hoch, JS .
STATISTICS IN MEDICINE, 2004, 23 (03) :411-429
[6]   THE CONCEPT OF RESIDUAL CONFOUNDING IN REGRESSION-MODELS AND SOME APPLICATIONS [J].
BECHER, H .
STATISTICS IN MEDICINE, 1992, 11 (13) :1747-1758
[7]   Controlling for continuous confounders in epidemiologic research [J].
Brenner, H ;
Blettner, M .
EPIDEMIOLOGY, 1997, 8 (04) :429-434
[8]   A potential pitfall in control of covariates in epidemiologic studies [J].
Brenner, H .
EPIDEMIOLOGY, 1998, 9 (01) :68-71
[9]   Parkinson's disease risks associated with cigarette smoking, alcohol consumption, and caffeine intake [J].
Checkoway, H ;
Powers, K ;
Smith-Weller, T ;
Franklin, GM ;
Longstreth, WT ;
Swanson, PD .
AMERICAN JOURNAL OF EPIDEMIOLOGY, 2002, 155 (08) :732-738
[10]   EFFECTIVENESS OF ADJUSTMENT BY SUBCLASSIFICATION IN REMOVING BIAS IN OBSERVATIONAL STUDIES [J].
COCHRAN, WG .
BIOMETRICS, 1968, 24 (02) :295-&