Table 2.
Likelihood ratio tests of logistic mixed model regressions on trout embryo survival within treatments
| Model | Effect tested | Model parameters | AIC | ln L | Likelihood ratio tests | |||
|---|---|---|---|---|---|---|---|---|
| Fixed | Random | δAIC | χ 2 | p | ||||
| a) Controls | ||||||||
| Reference model | A,Z | D,S | 386.2 | −188.1 | ||||
| Model 1 | A | Z | D,S | 387.6 | −189.8 | 1.4 | 3.1 | .07 |
| Model 3 | Z | A | D,S | 387.6 | −189.8 | 1.4 | 3.4 | .06 |
| Model 4 | D | A,Z | S | 426.1 | −209.1 | 39.9 | 41.8 | <.0001 |
| Model 5 | S | A,Z | D | 394.2 | −193.1 | 8 | 9.9 | .001 |
| Redness reference | 1 | D,S | 422.6 | −207.3 | ||||
| Model 6 | R | R | D,S | 420.7 | −207.3 | 1.9 | 0.09 | .8 |
| b) NB 1:1,000 | ||||||||
| Reference model | A,Z | D,S | 741.8 | −365.9 | ||||
| Model 1 | A | Z | D,S | 740.4 | −366.2 | 1.4 | 0.6 | .44 |
| Model 3 | Z | A | D,S | 741.9 | −366.9 | 0.1 | 2.1 | .14 |
| Model 4 | D | A,Z | S | 791.8 | −391.9 | 50 | 51.9 | <.0001 |
| Model 5 | S | A,Z | D | 746.2 | −369.1 | 4.4 | 6.4 | .01 |
| Redness reference | 1 | D,S | 753.9 | −372.9 | ||||
| Model 6 | R | R | D,S | 752.2 | −373.1 | 1.7 | 0.4 | .5 |
| c) NB 1:500 | ||||||||
| Reference model | A,Z | D,S | 1,126 | −557.9 | ||||
| Model 1 | A | Z | D,S | 1,128.9 | −560.4 | 2.9 | 4.1 | .04 |
| Model 3 | Z | A | D,S | 1,124.1 | −558 | 1.9 | 0.1 | .7 |
| Model 4 | D | A,Z | S | 1,144.6 | −568.3 | 18.6 | 20.6 | <.0001 |
| Model 5 | S | A,Z | D | 1,129.5 | −560.8 | 3.5 | 5.6 | .01 |
| Redness reference | 1 | D,S | 1,175.5 | −583.8 | ||||
| Model 6 | R | R | D,S | 1,174 | −583.9 | 1.5 | 0.4 | .5 |
Different logistic mixed‐effects models were compared to a reference (in bold) to test if the effects of astaxanthin (A), zeaxanthin (Z), dam (D), sire (S), and redness of the eggs (R) explain a significant part of the variance in embryo survival within (a) sham‐treated controls, (b) embryos exposed to low (NB 1:1,000) or (c) high nutrient broth concentrations (NB 1:500). Significant effects are highlighted.