Table 2.

Discrete choice parameter estimates (probit regression) based on responses from all countries

	N = 1775 (547 + 404 + 407 + 417)
	Obs = 35,500 [1775(j)×10(pairs)×2(I)]a
	Coef	SE	Sign
Constantb	−0.124	0.015	0.000
MO2	−0.299	0.031	0.000 0.000
MO3	−0.349	0.035
MO4	−0.923	0.036	0.000
MO5	−1.326	0.039	0.000
SC2	−0.208	0.033	0.000
SC3	−0.290	0.035	0.000
SC4	−0.793	0.036	0.000
SC5	−0.966	0.035	0.000
UA2	−0.194	0.032	0.000
UA3	−0.254	0.035	0.000
UA4	−0.769	0.035	0.000
UA5	−0.987	0.035	0.000
PD2	−0.248	0.033	0.000
PD3	−0.241	0.035	0.000
PD4	−1.017	0.036	0.000
PD5	−1.258	0.036	0.000
AD2	−0.195	0.034	0.008
AD3	−0.454	0.035	0.000
AD4	−1.183	0.037	0.000
AD5	−1.401	0.038	0.000
Log likelihood	−9043.843
Wald chi2 (20)	4817.43
AIC	18,129.686
BIC	18,307.709
Degrees of freedom		21

^a I Number of alternatives, j number of respondents

^bIn the set of coefficients the constant represents the alternative specific constant, capturing a tendency to always choose the first option