The output on the right shows the results for the
least squares analysis of variance and parameter estimates but without interaction.
We shall next run the model through the REML procedure (Stats → Mixed Models (REML)→
Linear Mixed Models) and compare with that obtained by least squares
analysis of variance.
A description of how the REML analysis can be conducted in R is illustrated in Mbunzi and Nagda (2009).
|
|
*** Regression ***
Response variate: WEANWT
Fitted terms: Constant + YEAR + SEX + AGEWEAN + DL + DQ + RAM_BRD +
EWE_BRD
***Estimates of parameters***
|
Estimate
|
s.e.
|
t(688)
|
tpr.
|
Constant |
12.95
|
1.07
|
0.26
|
0.797
|
YEAR
92 |
-1.566
|
0.293
|
-5.35
|
<.001
|
YEAR
93 |
-1.096
|
0.275
|
-3.98
|
<.001
|
YEAR
94 |
-2.833
|
0.358
|
-7.92
|
<.001
|
YEAR
95 |
-3.228
|
0.344
|
-9.39
|
<.001
|
YEAR
96 |
-2.351
|
0.390
|
-6.03
|
<.001
|
SEX
M |
0.478
|
0.169
|
2.82
|
0.005
|
AGEWEAN |
0.07022
|
0.00886
|
7.93
|
<.001
|
DL |
2.726
|
0.315
|
8.65
|
<.001
|
DQ |
-0.2689
|
0.0340
|
-7.91
|
<.001
|
RAM_BRD
R |
-0.443
|
0.173
|
-2.56
|
0.011
|
EWE_BRD
R |
-0.586
|
0.237
|
-2.48
|
0.014
|
***Accumulated analysis of variance ***
|
Change
|
d.f
|
s.s.
|
m.s.
|
v.r.
|
+ YEAR
|
5
|
1208.149
|
241.630
|
48.99
|
+ SEX
|
1
|
55.983
|
55.983
|
11.35
|
+ AGEWEAN
|
1
|
344.206
|
344.206
|
69.78
|
+ DL
|
1
|
151.513
|
151.513
|
30.72
|
+ DQ
|
1
|
275.795
|
275.795
|
55.19
|
+ RAM_BRD
|
1
|
44.881
|
44.881
|
9.10
|
+ EWE_BRD
|
1
|
30.223
|
30.223
|
6.13
|
Residual
|
688
|
3393.701
|
4.933
|
|
Total
|
699
|
5504.450
|
7.875
|
|
|
|