Questions to be addressed

By confining ourselves to just two yield measurements (dry corm weight and number of corms) recorded in Study 1 we shall

• determine the best month for planting each of the three landraces
• decide which one is the most suitable landrace to be grown by farmers for the formal market
• and determine the optimum level of fertiliser for growing this landrace.

We shall discover that delaying planting until January resulted in several plots with zero growth. We shall discuss how these missing values should be treated.

Finally, the design of Study 1 provides an example of a split-plot factorial experiment. We shall show

• how to analyse and report the results from such an experiment
• how to evaluate the basic assumptions for analysis of variance