Table 2.

Summary of the results of the statistical models used to analyze survival, oviposition success, and blood feeding success

Dep. Variable: Model (N) −2LL (Res)/AICa	Effectb	Fn:d/Z/Wc	P	Estimated
Survival: Mixed ANCOVA model (519), 8.2/12.2	Altitude (m)	67.21:26	0.0001	−0.0025/m
	Duration (h)	16.01:26	0.0005	−0.058/h
	Wind protection	6.11:26	0.021	−0.22 net vs. cloth
	Date [R]	1.5	0.067	0.21
	Residual [R]	3.7	0.0001	0.28
	Intercept	1001:8	0.0001	1.77
Survival (with species): Mixed ANCOVA model (344), 53.9/57.9	Altitude (m)	47.01:68	0.0001	−0.0026/m
	Duration (h)	1.41:68	0.0001	−0.071/h
	Wind protection	6.71:68	0.012	−0.23 net vs. cloth
	Species	1.12:68	0.35	0.065 S vs. M
	Date [R]	1.6	0.057	0.38
	Residual [R]	5.8	0.0001	0.07
	Intercept	1001:8	0.0001	1.97
Survival (with weather): Mixed ANCOVA model (519), 15.6/17.6	Altitude (m)	66.41:32	0.0001	−0.0025
	Duration (h)	21.81:32	0.0001	−0.049
	Wind cover	9.01:32	0.0053	−0.278
	RH	6.71:32	0.0145	0.0037
	Wind speed	18.61:32	0.0001	−00762
	Residual [R]	4	0.0001	0.029
	Intercept	1781:32	0.0001	1.91
Oviposition: Logistic regression (267) 361/368.5 Global Beta = 0: Wald = 9.13, P = 0.059	Altitude (m)	1.11	0.29	−0.0022 (0.99)
	Duration (h)	4.71	0.029	−0.10 (0.90)
	Wind protection	2.51	0.11	−0.27 (0.77)
	Intercept	3.91	048	1.21 (3.34
Egg batch size: GLM ANCOVA (121), Global model: F3:117 = 1.8 P = 0.16, R2 = 0.043	Altitude (m)	0.31:116	0.58	0.05
	Duration (h)	2.751:116	0.10	3.37
	Wind protection	2.651	0.11	21.4
	Intercept	3.761	0.055	54.4
Blood feeding: Logistic Regression (66) 88.4/97.3 Global Beta = 0: Wald = 2.03, P = 0.73	Altitude (m)	0.501	0.48	−0.003 (0.99)
	Duration (h)	0.121	0.73	0.041 (1.04)
	Wind protection	0.531	0.46	0.23 (1.26)
	Intercept	0.0021	0.48	0.06 (1.06)

^aDependent variable and the statistical model used in the analysis; N denotes the number of mosquitoes used in the model. The residual −2 log likelihood value is followed by the Akaike information criterion (AIC). For logistic regression analyses, we provide the global Wald χ ²-test and P-value testing the null hypothesis that all effects are zero. For GLM, we list global model test and R² values.

^bIndependent variables, with random variable followed by [R]. ‘Wind protection’ refers to covering the tube with net versus cloth (see text).

^c F-statistics with their corresponding numerator (n) and denominator (d) df for fixed effects and Z-statistics for random variables. The Wald χ ²-test for logistic regression is reported.

^dEstimate for categorical variables compare the two categories, e.g., the estimated survival of An. gambiae s.s. (S) was higher by 6.5% than that of An. coluzzii (M), although the difference was not significant.