Analysis of Covariance: Nonparametric

There are numerous benefits in adjusting an analysis for one or several covariates, but often these benefits come at the price of having to make additional assumptions, such as distributional assumptions or assumptions regarding the relationship between a covariate and a response variable. For example, the validity of the popular analysis of covariance model is predicated on a variety of assumptions. Clearly, the need to make assumptions, or to restrict the results so that they apply only when these assumptions are true, is a weakness, and it would be preferable to reap the benefits of covariate adjustment without having to rely on unreasonable assumptions. There is a body of nonparametric covariate adjustment techniques that aim to allow for this. Specific techniques depend on the nature of the covariate (e.g., binary or continuous). Keywords: analysis of covariance; stratified analysis; stratified design

[1]  P. Prorok,et al.  Concepts and problems in the evaluation of screening programs. , 1981, Journal of chronic diseases.

[2]  H Brenner,et al.  A Potential Pitfall in Control of Covariates in Epidemiologic Studies , 1998, Epidemiology.

[3]  J. Wittes,et al.  Analysis and interpretation of treatment effects in subgroups of patients in randomized clinical trials. , 1991, JAMA.

[4]  V W Berger,et al.  Detecting selection bias in randomized clinical trials. , 1999, Controlled clinical trials.

[5]  Michael D. Ernst,et al.  Parametric Analyses In Randomized Clinical Trials , 2002 .

[6]  G. Koch,et al.  A review of some statistical methods for covariance analysis of categorical data. , 1982, Biometrics.

[7]  V. Berger,et al.  Randomization Technique, Allocation Concealment, Masking, And Susceptibility Of Trials To Selection Bias , 2003 .

[8]  V W Berger,et al.  Pros and cons of permutation tests in clinical trials. , 2000, Statistics in medicine.

[9]  J. Robins,et al.  Adjusting for differential rates of prophylaxis therapy for PCP in high- versus low-dose AZT treatment arms in an AIDS randomized trial , 1994 .

[10]  V. Rahlfs,et al.  Scores: Ordinal Data with Few Categories—How Should They Be Analyzed? , 1993 .

[11]  V. Berger.  Improving the information content of categorical clinical trial endpoints. , 2002, Controlled clinical trials.

[12]  V. Berger,et al.  Adjusting for Ordinal Covariates by Inducing a Partial Ordering , 2004 .

[13]  G. Koch,et al.  Non-parametric covariance methods for incidence density analyses of time-to-event data from a randomized clinical trial and their complementary roles to proportional hazards regression. , 2000, Statistics in medicine.

[14]  Edgar Brunner,et al.  Nonparametric Hypotheses and Rank Statistics for Unbalanced Factorial Designs , 1997 .

[15]  J D Emerson,et al.  Analyzing data from ordered categories. , 1984, The New England journal of medicine.

[16]  Vance W Berger,et al.  Nonparametric Adjustment Techniques for Binary Covariates , 2005, Biometrical journal. Biometrische Zeitschrift.

[17]  P. Rosenbaum.  The Consequences of Adjustment for a Concomitant Variable that Has Been Affected by the Treatment , 1984 .

[18]  J. Pearl,et al.  Confounding and Collapsibility in Causal Inference , 1999 .

[19]  P. Lachenbruch,et al.  Anova, kruskal-wallis, normal scores and unequal variance , 1991 .

[20]  G G Koch,et al.  Issues for covariance analysis of dichotomous and ordered categorical data from randomized clinical trials and non-parametric strategies for addressing them. , 1998, Statistics in medicine.