Contrast Statement

In general linear models, a contrast statement defines a hypothesis to be tested concerning levels of one or more factors. For example, in an experiment testing a difference in the effect of a drug at 4 dosing levels, the ANOVA table tests for a difference in effect due to dose. A contrast analysis allows to further test the hypothesis that a difference in effect exists between dose level 1 and 2. This contrast statement is often expressed in the form of an equation, e.g., 1*MeanOfGroup1 - 1*MeanOfGroup1 + 0*MeanOfGroup3 + 0*MeanOfGroup4 = 0.