help estimates
-------------------------------------------------------------------------------
Title
[R] estimates -- Save and manipulate estimation results
Syntax
command reference
-----------------------------------------------------------------
Save and use results from disk
estimates save filename [R] estimates save
estimates use filename [R] estimates save
estimates describe using filename [R] estimates describe
estimates esample: ... [R] estimates save
-----------------------------------------------------------------
Store and restore estimates in memory
estimates store name [R] estimates store
estimates restore name [R] estimates store
estimates query [R] estimates store
estimates dir [R] estimates store
estimates drop namelist [R] estimates store
estimates clear [R] estimates store
-----------------------------------------------------------------
Set titles and notes
estimates title: text [R] estimates title
estimates title [R] estimates title
estimates notes: text [R] estimates notes
estimates notes [R] estimates notes
estimates notes list ... [R] estimates notes
estimates notes drop ... [R] estimates notes
-----------------------------------------------------------------
Report
estimates describe [name] [R] estimates describe
estimates replay [namelist] [R] estimates replay
-----------------------------------------------------------------
Tables and statistics
estimates table [namelist] [R] estimates table
estimates stats [namelist] [R] estimates stats
estimates for namelist: ... [R] estimates for
-----------------------------------------------------------------
Description
estimates allows you to store and manipulate estimation results:
o You can save estimation results in a file for use in later
sessions.
o You can store estimation results in memory so that you can
a. switch among separate estimation results and
b. form tables combining separate estimation results.
Remarks
estimates is for use after you have fit a model, be it with regress,
logistic, etc. You can use estimates after any estimation command,
whether it be an official estimation command of Stata or a user-written
one.
estimates has three separate but related capabilities:
1. You can save estimation results in a file on disk so that you can
use them later, even in a different Stata session.
2. You can store up to 300 estimation results in memory so that they
are at your fingertips.
3. You can make tables comparing any results you have stored in
memory.
Remarks are presented under the following headings:
Saving and using estimation results
Storing and restoring estimation results
Comparing estimation results
Saving and using estimation results
After you have fit a model, say, with regress, type
. sysuse auto
. regress mpg weight displ foreign
You can save the results in a file:
. estimates save basemodel
Later, say, in a different session, you can reload those results:
. estimates use basemodel
The situation is now nearly identical to what it was immediately after
you fit the model. You can replay estimation results:
. regress
You can perform tests:
. test foreign==0
And you can use any postestimation command or postestimation capability
of Stata. The only difference is that Stata no longer knows what the
estimation sample, e(sample) in Stata jargon was. When you reload the
estimation results, you might not even have the original data in memory.
That is okay. Stata will know to refuse to calculate anything that can
be calculated only on the original estimation sample.
If it is important that you use a postestimation command that can be used
only on the original estimation sample, there is a way you can do that.
You use the original data and then use estimates esample to tell Stata
what the original sample was.
See [R] estimates save for details.
Storing and restoring estimation results
Storing and restoring estimation results in memory is much like saving
them to disk. You type
. estimates store base
to save the current estimation results under the name base, and you type
. estimates restore base
to get them back later. You can find out what you have stored by typing
. estimates dir
Saving estimation results to disk is more permanent than storing them in
memory, so why would you want merely to store them? The answer is that,
once they are stored, you can use other estimates commands to produce
tables and reports from them.
See [R] estimates store for details about the estimates store and restore
commands.
Comparing estimation results
Let's say that you have done the following:
. sysuse auto
. regress mpg weight displ
. estimates store base
. regress mpg weight displ foreign
. estimates store alt
You can now get a table comparing the coefficients:
. estimates table base alt
estimates table can do much more; see [R] estimates table. Also see [R]
estimates stats. It works similarly to table but produces model
comparisons in terms of BIC and AIC.
Also see
Manual: [R] estimates
Help: [R] estimates save, [R] estimates describe, [R] estimates store,
[R] estimates title, [R] estimates notes, [R] estimates query,
[R] estimates dir, [R] estimates replay, [R] estimates table,
[R] estimates stats, [R] estimates for
