Table 5.

Random effect beta GLM of interval estimates predicted by causal valence, question types, and covariance in Study 1.

Variables	Parameter	Coefficient	SE	2.5%	97.5%
Random intercept		0.57	0.10	0.38	0.76
Location submodel
Intercept	b0	-1.06	0.13	-1.28	-0.80
Causal valence	b1	-0.21	0.10	-0.40	0.00
Covar2	b2	0.12	0.04	0.04	0.19
Covar3	b3	-0.11	0.04	-0.18	-0.03
Structure	b4	-0.10	0.04	-0.18	-0.02
Predictive	b5	0.28	0.05	0.19	0.38
Precision submodel
Intercept	d0	2.13	0.06	2.02	2.24
Causal valence	d1	0.04	0.08	-0.12	0.19
Covar2	d2	0.05	0.09	-0.11	0.22
Covar3	d3	-0.08	0.08	-0.24	0.07
Structure	d4	0.55	0.09	0.38	0.73
Predictive	d5	-0.74	0.10	-0.92	-0.54
Covar2 × Causal valence	d6	-0.15	0.06	-0.27	-0.03
Covar3 × Causal valence	d7	0.18	0.08	0.03	0.33

Causal valence is coded as -1 = generative condition and 1 = preventive condition. The two dummy variables for covariation conditions: Covar2 is coded as -1 = C1, 1 = C2, and 0 = C3 condition. Covar3 is coded as -1 = C1, 0 = C2, and 1 = C3 condition. The coefficients of Covar2/Covar3 represent the difference between the ratings in the C2/C3 condition and the overall mean/precision of the causal ratings. The two dummy variables for question type conditions: Structure is coded as -1 = strength judgment, 1 = structure judgment, and 0 = predictive judgment; Predictive is coded as -1 = strength judgment, 0 = structure judgment, and 1 = predictive judgment. The coefficients of structure/predictive represent the difference between the ratings in the structure/predictive question and the overall mean/precision of the causal ratings.