st: Debugging estimation changes

[Ed Gerrish <egerrish@gmail.com>]

This is a general question about debugging STATA changes.  Here is my dilemma:

About 8 months ago, I ran a set of models and got results.  I saved
them as my baseline results, wrote the paper, and worked on
robustness.  Recently, I noticed that results that should have been
similar to the baseline were not.  When I went back to run the
baseline, the results had changed.  I figured I made a coding error
and I did the following:

1. Checked my code for recent changes.
2. Uncompressed my archived do files and datasets and ran them - the
new results persisted.
3. Checked to make sure version control was on (it was).
4. Checked my custom ado's for updates.  There were some, but when run
on older versions, the new results were sustained.

Any ideas on how I should proceed?  It is unclear to me what my next
step is.  This certainly could be a STATA 11 vs. 12 problem -- but I
was hoping version control would prevent that from being true.

Thanks,

-Ed
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/