Importance and approaches of obtaining experimental uniformity
Abstract
119 words
The need for a high degree of experimental uniformity is clearly
shown by the fact in an experiment that results in a CV
(coefficient of variation) of 20%, 8 replications would be
required to detect the size of difference between treatments
that could be detected by 2 replications if the CV were 10%.
Realistically, only if the percentage difference between two
treatments is twice the magnitude of the CV, will the difference
be reliably detected at P<0.05. Detection of differences can
be enhanced by choosing the treatments carefully, using the
most effective experimental design, conducting the experiment
and gathering data with approaches that purposely minimize
non-treatment differences between units, and analyzing the
results with the best statistical approaches available.
Subjects
3 categoriesRelated papers
More in Working Document Series →- Row arrangement and nitrogen effects on growth yield of sorghum-legume intercrops
- The genetics of virulence of the Maize Streak Mastrevirus (MSV)
- Agronomic performance and genotype x environment interaction analysis of 12 elite potato genotypes in Uganda
- Genetic diversity of Phytophthora infestans in Uganda