|
Statistical modelling
The analysis of variance table is shown alongside. Both herd group and herd within herd group are statistically significant (P<0.001) (much higher mean squares than the residual mean square). The output shows that the grand mean is 347 kg.
If we then use Data → Calculations… to add this value to the residual values for each cow we can obtain the value (EXPTOT) that a cow would have been expected to have should all herd means have been equal. Thus, a cow from KANYAN, the region, where offtakes were on average the highest, will have its expected value (EXPTOT) adjusted downwards.
|
|
***** Analysis of variance *****
Variate: TOTYLD
| Source of variation |
d.f.
|
s.s.
|
m.s.
|
v.r.
|
| HDGRP |
7
|
1682692.
|
240385. |
31.46
|
| HDGRP.HERD |
29
|
1296565.
|
44709.
|
5.85
|
|
Residual |
350
|
2674420.
|
7641.
|
|
| Total |
386
|
5653677.
|
|
|
* MESSAGE: the following units have large residuals.
| *units* |
38
|
298.
|
s.e |
83.
|
| *units* |
81
|
-249.
|
s.e |
83.
|
| *units* |
159
|
292.
|
s.e |
83.
|
***** Tables of means *****
Variate: TOTYLD
Grand mean 347.
| HDGRP |
KANYAN
|
KASHON
|
KIKAAS
|
MASHA
|
| |
444.
|
373.
|
275.
|
343.
|
| rep. |
83
|
62
|
43
|
11
|
| HDGRP |
MUTONT
|
RUHENG
|
RUSHER
|
RYERU
|
| |
275.
|
248.
|
312.
|
363.
|
| rep. |
48
|
35
|
22
|
83
|
*** Standard errors of differences of means ***
| Table |
HDGRP
|
|
| rep. |
unequal
|
|
| d.f. |
350
|
|
| s.e.d. |
37.3X
|
min.rep
|
| |
28.0
|
max-min
|
| |
13.6
|
max.rep
|
|
|
