Bayesian Linear Regression Models: Bayes Factor

You can specify the model design for the analysis, including the approach that is used to estimate the Bayes factor for the Bayesian Linear Regression Models. The following options are available only when either the Estimate Bayes Factor or Use Both Methods Bayesian Analysis option is selected.

Null Model
When selected, the estimated Bayes factors are based on the null model. This is the default setting.
Full Model
When selected, the estimated Bayes factors are based on the full model and you can select variables to use and additional factors and covariates.
Variables
Lists all variables available for the full model.
Additional factor(s)
Select variables from the Variables list to use as additional factors.
Additional covariate(s)
Select variables from the Variables list to use as additional covariates.
Computation
Specify the approach to estimate Bayes factors. JZS method is the default setting.
JZS method
When selected, invokes the Zellner-Siow’s approach. This is the default setting.
Zellner's method
When selected, invokes the Zellner’s approach and you are required specify a single g prior value > 0 (there is no default value).
Hyper-Prior method
When selected, invokes the hyper-g approach and you are required to specify a shape parameter a0 for Inverse-Gamma distribution. You must specify a single value > 0 (the default value is 3).
Rouder's method
When selected, invokes the Rouder’s approach and you are required to specify a scale parameter b0 for Inverse-Gamma distribution. You must specify a single value > 0 (the default value is 1).