|
Study protocol/Statistical analysis
It is important to describe in detail how the data will be
analysed.
The three genotypes and three feeding regimes will be factorially arranged in a completely randomised, individually fed experiment. Each variable will be subjected to a least squares analysis of variance by GenStat. The model will be designed to determine the effect of genotype, feeding regime and their interaction on body weight, daily gain, nutrient utilisation, carcass traits and meat quality. The daily live weight gain for each animal will be calculated by linear regression of live weight on time.
The structure of the analysis of variance will be as follows:
| Source of variation |
d.f.
|
S.S.
|
M.S.
|
| Genotype |
2
|
| Feeding regime |
2
|
Genotype x Feed |
4
|
| Residual |
63
|
| Total |
71
|
When an F-test is significant the least squares means will be compared using Duncan's multiple range test. Variables measured as percentages will be transformed using the arc sine function.(Biometricians may question the use of Duncan's test
here, also possibly the need for a transformation. Two study questions address these issues.)
|
