Skip to main content

Table 4 Average posterior inference on sire (co)variances from ten replicates using the bivariate Student's-t (BSt) and normal (BN) fitted models with different residual degrees of freedom (DF).

From: Use of linear mixed models for genetic evaluation of gestation length and birth weight allowing for heavy-tailed residual effects

  

Fitted Model2

  

BS t

BN

True Parameters

True Model1

PM ± SE3

95% PPI4

ESS5

PM ± SE

95% PPI

ESS

= 2.0

BSt-4

2.24 ± 0.16

[1.32, 3.62]

20,983

2.31 ± 0.18

[1.30, 3.83]

20,694

 

BSt-12

2.44 ± 0.14

[1.48, 3.88]

24,767

2.39 ± 0.14

[1.44, 3.81]

25,794

 

BN-∞

2.49 ± 0.17

[1.52, 3.93]

26,715

2.49 ± 0.17

[1.53, 3.93]

28,154

= 1.5

BSt-4

1.41 ± 0.17

[0.40, 2.76]

23,326

1.43 ± 0.18

[0.33, 2.89]

23,146

 

BSt-12

1.93 ± 0.19

[0.83, 3.47]

27,428

1.89 ± 0.19

[0.80, 3.42]

28,683

 

BN-∞

1.77 ± 0.22

[0.72, 3.23]

28,441

1.77 ± 0.21

[0.71, 3.22]

30,064

= 4.0

BSt-4

4.23 ± 0.30

[2.59, 6.70]

24,192

4.33 ± 0.33

[2.57, 6.99]

23,978

 

BSt-12

4.77 ± 0.29

[2.97, 7.48]

27,674

4.80 ± 0.30

[2.98, 7.55]

29,102

 

BN-∞

4.46 ± 0.39

[2.79, 6.96]

29,013

4.45 ± 0.39

[2.78, 6.98]

29,365

  1. 1Used to simulate data
  2. 2Used in analysis of simulated data
  3. 3Posterior mean ± Standard Error
  4. 495% equal-tailed posterior probability interval based on the 2.5thand 97.5thpercentiles of the posterior density
  5. 5Effective sample size