Statistical modelling

Study A

Here we look at some analysis of data from Study A. Individual yields of Napier grass recorded in the randomised block experiment are contained in CS5Data1. Using Stats → Analysis of Variance… and choosing 'One-way ANOVA (in Randomized Blocks)' for the Design, and ticking %cv in the dialog box after clicking the Options… button, the analysis of variance below can be obtained for total dry matter yield.

We note first that that there is no variation across replicates (variance ratio = 0.85) and if we look at the bottom of the table we see that the coefficient of variation is 41.4%, implying a large variation across plots and indicating the difficulties in controlling the variation in a randomised block experiment that has to accommodate so many accessions.

Despite this there are significant differences in dry matter yields among the 53 accessions (P<0.001).