Table 3. Results of the ecological (EM) and the complete models (CM), and of the hierarchical partitioning applied to tawny owl roadkill data (Regression models – Coefficient: model coefficients of the explanatory variables, S.E.: standard errors, t-value: t test, p-value significance of the t test for the ecological and complete models; Hierarchical partitioning – I: independent contribution, J: joint contribution, Total: total contribution, I(%): percent independent contributions of individual variables for the explained variance of roadkill data, Z-score: statistical significance of independent contribution of variables, *p<0.05; see Table 1 for variable codes.

		Regression models				Hierarchical partitioning
	Variables	Coefficient	S.E.	t value	p-value	I	J	Total	I (%)	Z-score
EM	Intercept	−1.457	0.838	−1.739	0.086
	ROAD	3.129	0.727	4.301	<0.001	0.219	0.037	0.256	70.459	11.59*
	HABITAT	−0.836	0.927	−0.902	0.370	0.010	0.001	0.011	3.306	−0.180
	HQ5	0.659	0.292	2.257	0.027	0.081	0.030	0.112	26.235	3.65*
	Total					0.310
CM	Intercept	−0.326	0.672	−0.486	0.629
	ROAD	1.787	0.598	2.988	0.004	0.144	0.112	0.256	24.521	7.81*
	HABITAT	0.369	0.742	0.498	0.620	0.010	0.001	0.011	1.668	−0.180
	HQ5	0.310	0.233	1.332	0.187	0.057	0.055	0.112	9.614	2.38*
	SPA	0.958	0.139	6.887	<0.001	0.378	0.109	0.487	64.196	22.12*
	Total					0.589

(ecological model: AICc = 203.3, r = 0.557; complete model: AICc = 166.8, r = 0.767).