. 2016 Nov 29;9:616. doi: 10.1186/s13071-016-1905-7

Table 2.

Minimal adequate models for blood meal size, fecundity and survival

Categorical exposure level: no – low – high exposure						Quantitative continuous exposure level: bird parasitaemia
Lmer		Estimate	SE	t-value	P	Lmer		Estimate	SE	t-value	P
Blood meal size (square-root transformed)						Blood meal size (square-root transformed)
	Intercept	0.17	0.87	0.20	0.845		Intercept	0.74	0.99	0.75	0.455
	Body size	1.13	0.25	4.56	< 0.001		Body size	0.95	0.29	3.32	0.001
							Bird parasitaemia	0.14	0.06	2.40	0.016
R² _m = 3.4%; R² _c = 19.1%						R² _m = 4.5%; R² _c = 10.8%
Clutch size						Clutch size
	Intercept	-146.28	48.90	-2.99	0.003		Intercept	-147.60	54.87	-2.69	0.008
	Body size	74.58	14.58	5.24	< 0.001		Body size	66.28	16.15	4.10	< 0.001
	Blood meal size	2.76	0.43	6.36	< 0.001		Blood meal size	4.81	0.37	12.97	< 0.001
	Exposure level	-30.08	16.65	-1.82	0.076
	Blood meal : exposure	2.15	0.72	2.95	0.006
R² _m = 48.5%; R² _c = 50.6%						R² _m = 50.9%; R² _c = 52.5%
Coxme		Exp(coef.)	SE	z-value	P	Coxme		Exp(coef.)	SE	z-value	P
Survival						Survival
	Clutch size	1.01	0.001	10.29	< 0.001		Clutch size	1.01	0.001	9.83	< 0.001
Glmer		Estimate	SE	z-value	P	Glmer		Estimate	SE	z-value	P
Laying probability						Laying probability
	Intercept	-0.57	0.35	-1.60	0.110		Intercept	-0.56	0.33	-1.67	0.094
	Blood meal size	0.09	0.02	4.82	< 0.001		Blood meal size	0.09	0.02	4.21	< 0.001
	Exposure level				0.064		Bird parasitaemia	-0.45	0.16	-2.86	0.004
R² _m = 18.9%; R² _c = 26.4%						R² _m = 18.2%; R² _c = 18.2%

Each response variable (blood meal size, clutch size, survival and laying probability) was analysed with exposure level as a categorical variable (unexposed – low exposure – high exposure) and as a quantitative continuous variable (bird parasitaemia). Minimal adequate models, obtained after backward model selection based on likelihood ratio tests, are given with estimate or coefficient (exponential value, Exp(coef.)), standard error (SE), t- or z-value and P-value associated with each parameter in accordance with the performed models (lmer, coxme or glmer). Marginal and conditional R² (R² _m and R² _c, respectively) were calculated for mixed effects models according to [63]