Statistical modelling

Study A

Mean accession yields are shown in the table. These range from 21.0 kg for accession no. 15743 to 120.7 kg for accession no. 16803.

The five 'best bet' accessions selected for the feeding trial are shaded in green.

• Accessions 16803 and 168367 were selected from this experiment as the two showing the highest yields (see table).
• Accession no. 14984 was selected because it performed best in the two-site Ethiopian trial (Study B2). It can be seen that it performed well in this experiment too.
• Accessions 16786 and 16835 were selected as performing the best in the multi-location trial (Study B1). Note, however, how poorly accession no. 16835 performed in this experiment. This demonstrates the importance of not making conclusions from just one experiment, which, as shown on the previous page, had high plot-to-plot experimental variation.