. 2016 Feb 9;29(2-3):194–205. doi: 10.1002/bdm.1938

Table 1.

Mixed logistic regression analysis modeling choices in the partial and complete feedback conditions

Variables	Partial feedback				Complete feedback
Variables	Coeff	SE	z	p	Coeff	SE	z	p
Diff in expected values (dEV)	.1519	.0303	5.02	<.001	.1886	.0312	6.04	<.001
Anticipated disappointment (d)	.0225	.0098	2.29	.022	.0148	.0100	1.47	.140
Anticipated regret (r)	.0278	.0067	4.15	<.001	.0309	.0068	4.53	<.001
Interaction dEV × r	−.0033	.0023	−1.41	.159	−.0053	.0024	−2.17	.030
Interaction dEV × d	−.0045	.0026	−1.7	.088	−.0022	.0026	−.83	.408
Interaction d × r	.0007	.0004	1.72	.086	.0003	.0004	.64	.521
Interaction dEV × r × d	−.0001	.0002	−.4	.690	.0000	.0002	−.05	.961
Constant	.1785	.1109	1.61	.107	−.0025	.1092	−.02	.981
	Log likelihood = −501.75923				Log likelihood = −494.20136
	Wald χ ²(7) = 91.35				Wald χ ²(7) = 101.98
	Prob > χ ² = .0000				Prob > χ ² = .0000

The probability of choosing the left lottery over the right one is estimated as a function of the difference in expected values between the two lotteries, anticipated regret and disappointment. The regression includes the interactions of these three variables with one another.