ABSTRACT
The regression models discussed in the previous chapters tted straight line relationships between a response variable and one or more explanatory variates. In many situations, this type of model adequately reects the observed pattern, but sometimes a curved relationship is observed, or there might be a biological or physical reason for a curved relationship. Fitting a straight line will then produce an inadequate model and an alternative approach should be sought. In this chapter, we consider some simple techniques for tting curved relationships in terms of one or more explanatory variates.
