Table 3. The effect of being recipient of a good action on altruism in the Dictator game. We used logistic regression with and without control on sex, age, and education, using ‘treat’ as a dummy variable. We report the β-value, its standard error, and the significance level. ‘Bas’ stands for ‘Baseline’, ‘Con’ for ‘Control’, and ‘Exc’ for ‘Excluded’. So, for instance, the column ‘Bas Exc vs Con 1 Exc’ reports the results of the regression using Bas Exc = 0 and Con 1 Exc = 1 as dummy variable. We find no significant effect of the dummy on donations in the DG. Being recipient of a good action does not significantly increase the probability of splitting the endowment in the DG.

	Bas vs Con 1	Bas vs Con 1	Bas Exc vs Con 1 Exc	Bas Exc vs Con 1 Exc	Bas vs Con 2	Bas vs Con 2	Bas Exc vs Con 2 Exc	Bas Exc vs Con 2 Exc
dec
treat	−.082	−.115	.068	.024	.163	.257	.186	.289
	(.27)	(.27)	(.28)	(.28)	(.27)	(.28)	(.27)	(.29)
sex		.545		.436		.624*		.619*
		(.28)		(.29)		(.31)		(.31)
age		.030		.028		.015		.015
		(.02)		(.02)		(.02)		(.02)
education		.047		−.006		−.194		−.229
		(.11)		(.12)		(.11)		(.12)
Constant	−.251	−2.078**	−.186	−1.556*	−.251	−.746	−.186	−.520
	(.19)	(.74)	(.19)	(.76)	(.19)	(.74)	(.19)	(.77)
Pseudo R2	.000	.032	.000	.023	.001	.032	.002	.035
No. of cases	227	227	210	210	225	225	214	214

Significance level:

**: p < 0.01.

*: p < 0.05.