Table 2.
Likelihood-ratio tests of variance components
| Modela | No. | Test | log(L) | Δlog(L) | P | Significance of… |
|---|---|---|---|---|---|---|
| Tarsus length | ||||||
| E | 1 | – | −56.21 | |||
| E Q | 2 | 2 versus 1 | −56.21 | 0 | – | Experimental quartet effect |
| E R | 3 | 3 versus 1 | 45.32 | 101.52 | <0.001 | Nest-of-rearing effect |
| E R O | 4 | 4 versus 3 | 69.67 | 24.35 | <0.001 | Nest-of-origin effect |
| E R O A | 5 | 5 versus 4 | 72.55 | 2.88 | 0.008 | Additive genetic effect (V A) |
| E R O Age(A) | 6 | 6 versus 5 | 73.14 | 0.58 | 0.146 | Age dependence of V A |
| Age(E) R O A | 7 | 7 versus 5 | 73.34 | 0.79 | 0.103 | Age dependence of residual variance (V E) |
| Age(E) R O Age(A) | 8 | 8 versus 7 | 73.35 | 0.01 | 0.499 | Test for confounding effect of V E on V A |
| 8 versus 6 | 73.35 | 0.78 | 0.103 | |||
| Body mass | ||||||
| E | 1 | – | −302.24 | |||
| E Q | 2 | 2 versus 1 | −300.09 | 2.14 | 0.038 | Experimental quartet effect |
| E Q R | 3 | 3 versus 2 | −156.61 | 143.47 | <0.001 | Nest-of-rearing effect |
| E Q R O | 4 | 4 versus 3 | −134.18 | 22.43 | <0.001 | Nest-of-origin effect |
| E Q R O A | 5 | 5 versus 4 | −133.27 | 0.91 | 0.061 | Additive genetic effect (V A) |
| PHA response | ||||||
| E | 1 | – | 360.21 | |||
| E Q | 2 | 2 versus 1 | 367.87 | 7.61 | <0.001 | Experimental quartet effect |
| E Q R | 3 | 3 versus 2 | 415.13 | 47.3 | <0.001 | Nest-of-rearing effect (V R) |
| E Q R O | 4 | 4 versus 3 | 430.06 | 14.94 | <0.001 | Nest-of-origin effect (V O) |
| E Q R O A | 5 | 5 versus 4 | 431.64 | 1.57 | 0.003 | Additive genetic effect (V A) |
| E Q R O Age(A) | 6 | 6 versus 5 | 447.96 | 16.33 | <0.001 | Age dependence of V A |
| Age(E) Q R O A | 7 | 7 versus 5 | 431.65 | 0.01 | 0.499 | Age dependence of residual variance (V E) |
| Age(E) Q R O Age(A) | 8 | 8 versus 6 | 448 | 0.04 | 0.479 | Test for confounding effect of V E on V A |
| 8 versus 7 | 448 | 16.35 | <0.001 | Age dependence of V A in presence of age-dependent V E | ||
| E Q R O Age(A)b | 9 | 9 versus 6 | 448.02 | 0.06 | 0.485 | Cross-age genetic covariance lower than unity |
| E Q Age(R) Age(O) Age(A) | 10 | 10 versus 6 | 449.09 | 2.77 | 0.09 | Test for confounding effect of V R and V O on V A |
| E Q Age(R) O Age(A) | 11 | 11 versus 6 | 449.08 | 2.76 | 0.09 | Test for confounding effect of V R on V A |
| E Q R Age(O) Age(A) | 12 | 12 versus 6 | 447.95 | 0.01 | 0.99 | Test for confounding effect of V O on V A |
Bold indicates significant results in model comparisons
log(L), logarithm of likelihood; Δlog(L), difference in log-likelihoods of the more complex and simpler model; P, significance of the random effect added in the more complex model, as compared to the simpler model; Test, which models were compared. The last column provides the interpretation of each model comparison
aTerms in models are labelled in the following way: E, residual variance; Q, quartet; R, nest of rearing; O, nest of origin; A, additive genetic effect; Age(X), (constrained) age-dependent covariance matrix is fitted (cross-age correlations constrained to unity for A and zero for E)
bResulting covariance matrix is unconstrained (covariance is estimated)