ABSTRACT
This chapter focuses on additional inferential techniques applied to contrasts of interest. It considers post analysis of variance f–test procedures for carrying out treatment comparisons planned in advance of the experimental results. Planned comparisons are those comparisons which are suggested by the original goals of the experiment. It is best for a user to plan in advance which treatment comparisons are of interest and to find simultaneous confidence intervals only for these interesting comparisons. In some problems certain comparisons become interesting or logical to consider after an initial analysis of the data has been performed. When a treatment factor in an experiment is of the quantitative type, the experimenter sometimes finds it useful to fit a response curve in order to determine the mathematical nature of the response to levels of the factor.
