Data Set 6: Effects of Host Species and Size on Parasitoid Size

This handout illustrates two-factor ANOVA with both factors being fixed effects. Two versions of essentially the same data set are used: (A) a completely balanced data set, i.e. one in which there is equal replication of each treatment combination [this is a sub-set of the actual data set]; and (B) An unbalanced data set, with unequal replication of treatments but all treatment combinations present [this data set was artificially constructed but mimics the actual treatment effects]. The simpler balanced situation is used to illustrates the analysis of the three overall effects and comparisons among levels of factors; assessment of the validity of the ANOVA model also is briefly discussed for this data set. The second version illustrates the complications produced by unbalanced designs, and how these are dealt with. At the end of this handout are a detailed description of how to conduct a two-way fixed-effects ANOVA in Minitab, and a sketch of how to do so in SAS.