Reporting
Another way to present the results is to use the GenStat
Predict command to produce least squares estimates of means for each
factor. This command becomes available within the Genstat dialog box once the model has
been fitted. Note that this needs to be done for each factor one at a time; when
all three factors are predicted simultaneously a 3-way table of least squares
means is produced.
It is helpful to include the numbers of observations for each
level of each parameter, both in this table and sometimes in the previous one
too. These can be produced using the Stats → Summary Statistics → Frequency Tables...
command. It can be seen from the table that the number of
observations influences the magnitudes of the standard error, the larger the
number of observations generally the smaller the standard errors.
Regression coefficients have
also been included at the foot of the table. This is optional.
|
|
|
Parameter
Genotype |
No. |
Mean
|
S.E.
|
|
Dorper (D) |
220 |
11.56 |
0.16 |
Red Maasai (R) |
181 |
10.51 |
0.19 |
R x D |
198 |
11.16 |
0.18 |
D x R |
101 |
11.04 |
0.24 |
Year |
|
|
|
1991 |
144 |
12.73 |
0.22 |
1992 |
109 |
11.16 |
0.22 |
1993 |
168 |
11.63 |
0.18 |
1994 |
79 |
9.91 |
0.26 |
1995 |
107 |
9.53 |
0.24 |
1996 |
93 |
10.39 |
0.28 |
Sex |
|
|
|
Female |
323 |
10.84 |
0.12 |
Male |
377 |
11.32 |
0.12 |
|
|
Regression coefficient |
S.E.
|
Age at weaning (day) |
|
0.070 |
0.010 |
Age of dam (year) |
|
|
|
Linear |
|
2.19 |
0.25 |
Quadratic |
|
-0.269 |
0.034 |
|
|