1.5 Over the last two weeks we have been considering experimental designs and statisticalmodels that only include one ‘type’ of predictor variable. But this week is different. Now weare looking at models that include two ‘types’ of predictor variables. Write down the wordequation for this analysis, a) including the interaction, and b) excluding the interaction. Makedefine all terms.

1.6 If we fit model excluding the interaction, what are we assuming about the effect of size onseeds?

1.4 Nevertheless, you have measurements on subsamples. How can you include this informationin your analysis while avoiding treating these subsamples as true replicates?

We hope that you have correctly identified the danger of pseudoreplication in this exam-ple. It is important to recognize that there is not necessarily a problem with taking measure-ments on subsamples. However, there is a problem if you include these measurements as truereplicates in your analysis. What would be the consequence of including measurements onsubsamples as true replicates in your analysis (of covariance)?

1.2 How many independent replicates are there within each level of the experimental treat-ment?

1.1 Draw out the experimental design, indicating measurement and experimental units and alsothe total number of experimental units (true replicates) and the total number of measurementunits