Comparison between complete block design, one dimensional alpha-lattice and post-blocking analyses in small grain variety testing in Spain
2002
El-Jiati, A.
The objective was to compare different statistical models to improve precision when comparing varieties. We compared the effectiveness of different models: randomised complete block (RCBD) vs. one-dimension randomised incomplete block (RIBD); and row-column post-blocking (RCPB). We evaluated the gains in efficiencies attained by new analytical methods, namely one- and two-dimensional spatial analysis. Use of RCPB, alpha-lattice, improved the efficiency to compare variety estimates. Post-blocking introduction of column on top of regular one-dimensional row alpha designs, considerably increased the efficiency of the analysis. However, when analysing a single near-optimum design in a different crop, the use of one and two-dimensional spatial methods did not achieved the levels of precision of the conceptually simpler two-dimensional incomplete block design. Furthermore, some difficulties still exist: the difficulties associated with both the structure of the computer programs to be used, and the need to identify the best spatial method for a given field trial.
显示更多 [+] 显示较少 [-]