Bias approximations for complex estimators: An application to redundancy analysis.

Presents and illustrates a method for approximating the amounts of bias in estimators having complex sampling distributions that are influenced by a variety of properties typical of applied research data. The method, which appears to have broad applicability to many multivariate procedures, is illustrated in the context of redundancy analysis. Results.