Stata tip 116: Where did my p-values go? (part 3)
The Stata Journal (2014) Vol. 14 No. 1, pp. 218-220.
Abstract
In a previous Stata tip I discussed how to recover t-statistics, p-values, and confidence intervals for regression parameters using the results that are returned by an estimation command. In a subsequent Stata tip I discussed how to recover parameter estimates for parameters that were estimated on a transformed scale, for example if a likelihood function contains a standard deviation or a correlation than many Stata commands will maximize the likelihood with respect to ln(standard deviation) and the Fisher's z-transformation of the correlation. In this tip I will discuss how to recover the standard errors for the back-transformed parameters, that is, the standard error for the standard deviation and the correlation. I will also discuss how to retrieve the confidence intervals as returned by Stata.