Table 2. Logistic model outcomes describing the odds of early maturation (age 2 in males and age 3 in females) for Glenariffe versus Hakataramea contrasts.

Model	Glenariffe versus Hakataramea Terms (P-values and odds effects)	N-R2	AIC	ΔAIC
Males
Population	Pop<0.001	0.017	1232.87	495.15
Full model	Pop <0.001, F12<0.001, F16=0.131, F19=0.053, W12=0.003, W16=0.929, W19=0.490, GF16<0.001, GF19=0.023, GW16=0.001, GW19=0.003	0.534	739.55	1.83
Backwards+populationa	Pop=0.018, F12<0.001, F19<0.001, W12<0.001, GF16<0.001, GF19<0.001, GW16<0.001, GW19<0.001	0.531	737.72	—
Backwards−population	FL12<0.001, FL19<0.001, W12<0.001, GF16<0.001, GF19<0.001, GW16<0.001, GW19<0.001	0.523	741.39	3.67
Females
Population	Pop<0.001	0.015	1356.53	1189.25
Full model	Pop=0.133, F24=0.496, F28=0.873, F31=0.460, W24=0.633, W28=0.210, W31=0.118, GF28=0.391, GF31=0.467, GW28=0.238, GW31=0.001	0.940	176.48	9.20
Backwards+population	Pop=0.204, W28=0.002, GF28=0.002, GW28<0.001, GW31<0.001	0.939	167.64	0.36
Backwards−populationa	W28=0.003, GF28=0.003, GW28<0.001, GW31<0.001	0.938	167.28	—

Abbreviations: AIC, Aikiake's information criterion; Pop, population; F no. or W no., fork length or weight at the specified month number; GF no. or GW no., growth in fork length or growth in weight during the interval leading up to the specified month number; N-R², Nagelkerke's R².

Model terms, P-values, N-R², AIC and deviations in AIC (that is, ΔAIC) from the model with lowest AIC score are presented for each model and sex. Backwards stepwise models are depicted with and without population effects to better assess contribution to model performance. Model terms depicted in bold had positive marginal effects on odds of early maturation, those presented in italics had negative marginal effects on early maturation. Population effect reflects influence of a fish being of Glenariffe origin as opposed to Hakataramea origin.

^aMost parsimonious model identified under backwards stepwise regression.