Table 2. Parameter estimates weight-averaged across the 51 “best” models predicting Andean potato weevil infestations (square root of proportion infested tubers).

			95% CI
Variable	Estimate	SEM	Lower	Upper	ω p
Intercept	1.774	0.936	−0.061	3.608	1
Observer effect 1	−0.043	0.019	−0.081	−0.005	1
Observer effect 2	−0.070	0.016	−0.102	−0.038	1
Perimeter/area ratio	0.519	0.150	0.225	0.814	1
Neighboring storage units	0.044	0.015	0.015	0.073	1
Neighboring current potato	−0.004	0.002	−0.007	−0.001	1
Insecticide treatments January	−0.053	0.021	−0.094	−0.013	1
Insecticide treatments December	−0.054	0.025	−0.103	−0.004	1
Neighboring previous potato	0.007	0.004	0.000	0.014	0.91
Soil clay	0.005	0.003	0.000	0.010	0.91
Elevation (sqrt transformed)	−0.031	0.013	−0.057	−0.005	0.83
Number of hillings	0.127	0.076	−0.022	0.275	0.71
Rotation 2007 [Not potato]	−0.096	0.062	−0.216	0.025	0.71
Rotation 2006 [Not potato]	0.030	0.020	−0.008	0.069	0.49
Height of first potato hilling	−0.006	0.004	−0.014	0.002	0.43
Height of row at harvest	−0.005	0.003	−0.012	0.001	0.42
Chemical fertilization at planting	−0.002	0.002	−0.006	0.002	0.23
Soil organic matter	−0.012	0.008	−0.026	0.003	0.21

Notes: Estimates are followed by their standard errors, their 95% confidence intervals and their Akaike parameter weights (ω _p). Given a set of similarly-adequate predictive models, parameter weights estimate the probability that the parameter is included in the best model in the set.