Statistical Theory and Methods
Published Online: 15 SEP 2006
Copyright © 2002 John Wiley & Sons, Ltd
Encyclopedia of Environmetrics
How to Cite
Hampel, F. 2006. Robust Inference. Encyclopedia of Environmetrics. 5.
- Published Online: 15 SEP 2006
Robust inference is inference that is insensitive to (smaller or larger) deviations from the assumptions under which it is derived. Some very commonly used assumptions in statistics are normality, independence, identical distributions, linearity (e.g. in regression), and stationarity of stochastic processes. As a rule, such assumptions are only approximations to reality, and the questions arise as to what deviations tend to occur in practice, what effects they have on known statistical procedures, and how to develop better, ‘more robust’ procedures when deemed desirable. To answer these questions, new statistical concepts have proven useful.