Table 2. Mixed Logit model results for the general and local samples.

	General sample				Local sample
Parameter	Estimate1	Standard Error	t Value	P	Estimate1	Standard Error	t Value	P
Landsc. rich mosaic	0.3982	0.2171	1.83	0.0666	0.7428	0.2765	2.69	0.0072
Landsc. abandonment	−1.0471	0.3066	−3.41	0.0006	−0.7978	0.2873	−2.78	0.0055
Biodiversity 15	0.8877	0.3069	2.89	0.0038	0.2609	0.2144	1.22	0.2237
Biodiversity 7	−0.8434	0.2947	−2.86	0.0042	−0.5034	0.2406	−2.09	0.0364
Forest fires 6	−2.8342	0.9871	−2.87	0.0041	−1.4563	0.5573	−2.61	0.0090
Forest fires 2	2.5707	0.8265	3.11	0.0019	1.1894	0.3797	3.13	0.0017
Prod. quality 6	0.9789	0.4158	2.35	0.0186	0.7589	0.3405	2.23	0.0258
Prod. quality 2	−2.0904	0.7382	−2.83	0.0046	−1.1044	0.4721	−2.34	0.0193
Annual cost	−0.0399	0.0121	−3.30	0.0010	−0.0150	0.0082	−1.81	0.0697
Model fit
No. respondents	402				102
No. obs.	2010				510
Log likelihood	−1892				−480.36
McFadden LRI	0.1434				0.1427

¹estimated regression coefficients that express the “marginal utility” of each attribute level.