Examining the posterior distributions for current and beginning salaries

While still considering the logarithm of the current and beginning salaries for the 474 employee records we have observed and log-transformed, in a more practical scenario, assume that neither of the salary variance values are known. In this scenario, the small business owner wants to make Bayesian inference on the means of the two log-transformed salary variables by placing a scaled inverse Chi-square distribution on the variances, and a normal distribution on the means conditional on the variances. A parameter to scale the variances is also specified.

This information is collected in Employee data.sav. See the topic Sample Files for more information.

Next