Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: RE: new graphing command -twoway estfit- available on SSC

From   "Nick Cox" <[email protected]>
To   <[email protected]>
Subject   st: RE: new graphing command -twoway estfit- available on SSC
Date   Fri, 18 Jun 2004 10:35:01 +0100

This is a very welcome new command. For those 
interested, there is some overlap in functionality 
with -regplot- on SSC. The main differences are that 

1 -regplot- is a post-estimation graphing command intended
to work when fitted response is a smooth function of 
whatever is on the x axis (by default, the first-named 
predictor). -reg- does _not_ denote limitation 
to -regress-. 

2 as Michael emphasises -twoway estfit- is a -twoway- 
type and meshes with other -twoway- commands. That's 
a very important feature. -regplot- is stand-alone at 
present but I'll pester the author to add a -plot()- 

[email protected] 

Michael Blasnik
> I would like to tell Statalist that a new graphic command that I've
> "written" has been posted to SSC (thanks to Kit Baum).  The new graph
> command is -twoway estfit- as in :
> graph twoway estfit y x , .....
> -twoway estfit- plots the fitted values from any estimation 
> command (well,
> any command that works with -predict- and just needs an x/y 
> pair).  It is
> simply a modified version of -twoway lfit-, which uses 
> -regress- only, and
> other similar fitted value plot commands such qfit, fpfit, etc..
> -twoway estfit- has an option -estcmd(estimation command 
> name)- which lets
> you tell it what estimation command to use.  For example, you 
> can specify
> using -rreg- or -qreg- to get more robust linear fits than 
> -lfit- might
> provide, or even use it with -logit- or -probit- to show the fitted
> relationships.  Although you can get the same effect by
> just running your estimation command, running predict and 
> then making a line
> graph of the fitted values, by making it a full twoway graph 
> type you get
> access to the power of all of Stata's twoway options, by() is 
> especially
> useful.  I find myself using this graph type frequently.
> This was my first foray into .class programming and all the 
> new graphics
> stuff in Stata, but it wasn't too hard since I basically 
> lifted everything
> from -lfit-.   I haven't yet tackled -estfitci-, but that 
> should be doable
> too if there's interest.
> You can find it by -ssc describe t- where twoway_estfit will 
> be listed.

*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index