ABSTRACT
The box-plot techniques discussed in the previous chapter provide a con venient way to compare continuous response variate densities graphically. These plotting techniques correspond to a class of analytic techniques for the study of experimental and observational (Ci ^ 2 ) data known as analysis of variance (ANOVA) that has a comprehensive literature made up of many fine books and important papers. The analysis of this type of information is facilitated by a quantity called a noncentrality parameter δ. However, before explaining what δ actually is-in terms of the (C1 I.A2 ) data case where Λ 2 assumes three or more values-the t-test for group differences, which concerns the special case of (C1 IA2 ) where Λ 2 assumes two distinct values, will provide relevant background. Assume that the C\ variate has a Normal density and the binary, in other words dichotomous-valued vari able Λ2 determines the index j of the location parameter μj of this Normal density, an index that takes on the value 1 or 2 .
