Stata 10 allows you to store estimation results in a file for use in later sessions. Fit a model today, save the results to disk, come back tomorrow, load the results, and continue with postestimation analysis as if you never exited Stata. Or email your results file to a colleague for further review. estimates save and estimates use make these tasks easy. Here is an example:
First, we fit a multinomial logit model:
. use http://www.stata-press.com/data/r10/sysdsn3 . mlogit insure age male nonwhite site2 site3
We are about to leave for the day, so we’ll save our mlogit results to disk so tomorrow we can pick up where we left off:
. estimates save mymlogitresults file mymlogitresults.ster saved
Now we’re back and ready to get back to work. We load in the estimation results from yesterday:
. estimates use mymlogitresults
and now we can do our postestimation analysis. Here we replay our results and then test the joint significance of the coefficients in the Prepaid equation:
We did not reload the dataset when we came back to our work. estimates use does not require that the dataset used during estimation be available. Since our hypothesis test is based only on the estimated parameters and their covariance matrix, the dataset is not needed. If you want to make predictions or describe the estimation sample, then you do of course need the original dataset, and you use the estimates esample: command after estimates use to define the estimation sample.
For a complete list of new data management features, click here.