Changes to outreg since version 3.
This version is a complete rewrite of outreg, mostly programmed in Mata. All it shares with the previous versions of outreg other than the name is part of the command syntax.
The objective of this version was as complete control of the layout and formatting of estimation tables as possible, creating both Word and TeX files.
1. The most obvious change to outreg is that it now writes fully formatted Word or TeX files rather than flat text files. For this reason, there are a lot of new options relating to fonts, text justification, cell border lines, etc. that were not needed before.
2. outreg can now display any number of statistics for each estimated coefficient (instead of just two), and these statistics can be displayed side-by-side as well as one above the other. One can choose from 26 different statistics for inclusion in a table, including the full panoply of marginal effects and confidence intervals.
3. One can now selectively keep or drop coefficients or equations from the estimation results.
4. Tables can be merged or appended more flexibly than before. Multiple tables can also be written successively to the same file with regular paragraphs of text in between, so that it is possible to create a whole statistical appendix in a single document with a .do file.
5. Text can include italic, bold, super and subscripts, and Greek characters. It is possible to include user-specified fonts ( addfont). Column titles can span multiple cells (multicol). Footnotes can be added to any part of the table (annotate).
6. The table created by outreg is displayed in the Stata results window, minus some of the finer formatting destined for the Word or TeX file.
Some outreg syntax changes may cause confusion:
1. Name of "append" option changed. Successive estimation results are now combined with the merge option, which was named the "append" option in previous versions of outreg. This makes the new outreg consistent with the way the Stata merge command works on datasets versus the Stata append command.
2. By default, variable labels are not used. In the new outreg, variable labels replace variable names only when the varlabels option is chosen.
2. By default, multiequation models are not merged into multiple columns. Estimated coefficients from multiequation models like reg3 and mlogit are reported in one (long) column by default. To merge them into