Beta Regression on Causal Ratings Predicted by Evidence Types and Contingency Conditions.

The dummy variable C2 was coded as 1 for the zero contingency condition, and 0 for the other two contingency conditions; C3 was coded as 1 for the negative contingency condition, and 0 for the other two contingency conditions. The dummy variable Unambiguity was coded as 1 for the unambiguous condition, and 0 for the other two conditions; Omitting was coded as 1 for the omitting condition, and 0 for the other two conditions. Parameter b are the parameters of the variable in the location submodel, while parameter d are the parameters of the variable in the precision submodel (see S2 Appendix for a more detailed description). A positive b value indicates that the mean of the condition coded as 1 is higher than the condition coded as 0. A positive d value indicates that the precision of the condition coded as 1 is greater than the condition coded as 0. 2.5%CI and 97.5%CI are the lower and upper bounds of the 95% Bayesian credibility interval. Beta Regression on Causal Ratings Predicted by Evidence Types and Contingency Conditions.