Question1

Q1  Once I have the analysis for an experiment and I am particularly interested in a pre-planned comparison of treatments how can I use contrasts?

So we assume from this someone has designed, conducted and is analysing something like a One-way ANOVA or a factorial ANOVA typically with blocking. So how can we compare certain treatments with others or a group of treatments with another.

Example would be compare a genotype  A with the mean of B, C and D.

To do this we need to make an orthogonal contrast, and the Demonstration of this is downloadable here (Contrast)

To interpret this – we consider our contrast to be asking a particular question about the treatment structure and we are in fact dividing up our degrees of freedom for that factor into a particular question.

(Add an example R code and output)

Relevant concepts:     Levels of a Factor, orthogonal contrasts, contrast matrix