Robust regression: An introduction

Abstract

Analytical scientists use regression methods in two main areas. Calibration graphs are used with the results of instrumental analyses to obtain concentrations from test samples. Graphical methods are used to evaluate the results obtained when two methods, often a novel one and a reference one, are compared by applying them to the same set of test materials. In either case outliers or suspect results may occur, and exert big effects on the plotted regression line and the results derived from it. Robust methods are well suited to tackling such situations. Here some of the underlying ideas are summarised: later briefs will describe some more of the many approaches available.