The
accuracy of the reference – calculations
Marjatta
Oksama
Results
from dissection trials carried out in Sweden, Spain, Belgium and Netherlands.
The
model, we have used, is a variance component model, which allows dependence
between observations on the same carcass. With this model we can explain all
systematic differences between left and right side, slaughterhouses and fat
classes.
Model:
a
»
Country i = 14
fixed
J
»
Fatgroup j = 13
fixed
g
»
Abattoir k = 18
fixed
V
»
Sex l = 13
fixed
x
»
Side m = 1,2
fixed
B »
Butcher n = 18
random
C » Carcass o = 132 random
e »
error
random
All
random effects are assumed to be normally distributed with mean 0 and variance s^{2}.
Definitions:
Repeatability
standard deviation s_{r}
Reproducibility
standard deviation s_{R}
Reliability
=
Dissection lean meat percent: estimated butcher and carcassvariation, repeatability,
reproducibility and reliability.

s_{B} 
s_{C} 
s_{r} 
s_{R} 
Reliability 
Meat% 
0.68 
2.81 
0.87 
1.10 
0.87 