Table 3.

Random-effect tobit regression for σγ and ργ in each type of game.

	Standard dictator (θ <0)		Only recipient (θ = 0)		Efficient dictator (θ > 0)		Only dictator (θ = ∞)
	σ (γ)	σ (γ)	σ (γ)	σ (γ)	σ (γ)	σ (γ)	ρ (γ)	ρ (γ)
Reflecitve (R)	−0.719***	−0.651***	1.759***	1.929***	3.103***	2.926***	2.296***	2.137***
	(0.10)	(0.10)	(0.25)	(0.27)	(0.53)	(0.52)	(0.33)	(0.30)
Impulsive (I)	−0.454***	−0.333***	1.154***	1.462***	3.003***	2.686***	1.889***	1.602***
	(0.09)	(0.10)	(0.22)	(0.26)	(0.51)	(0.51)	(0.25)	(0.22)
Others	−0.535***	−0.445***	1.310***	1.614***	2.469***	2.148***	2.211***	1.980***
	(0.15)	(0.15)	(0.40)	(0.42)	(0.59)	(0.59)	(0.43)	(0.41)
Round	−0.003	−0.003	−0.002	−0.001	0.001	0.004	0.017	0.016
	(0.01)	(0.01)	(0.01)	(0.01)	(0.02)	(0.02)	(0.01)	(0.01)
Female		−0.179**		−0.507**		0.448		0.476***
		(0.07)		(0.20)		(0.29)		(0.18)
Wald test (R) = (I)	3.48***	3.99***	2.87***	2.19**	0.35	0.79	2.15**	2.65***
Observations	960	960	480	480	384	384	480	480

The Wald test after each regression is used to test the behavioral hypothesis that reflective and impulsive subjects behave in the same manner. Significance at the

^***1%,

^**5%,

^*10% level.