Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: st: referencing eresult matrices


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   RE: st: referencing eresult matrices
Date   Fri, 6 Nov 2009 13:36:13 -0000

I don't know one. I guess the low-level syntax parser needs a subscript to be preceded by a matrix or variable name and tweaking it so this could be understood is more bother than it's worth. As you know, you have to worry about what such a change might break. 

Nick 
n.j.cox@durham.ac.uk 

Jeph Herrin

Thanks, but as the subject line says, I want to
reference e(matrix) directly - I just gave one
example. I've been using the below method for years,
so I hardly think about it anymore, but every once
in a while wonder if there's a shorter way.

Maarten buis wrote:

> --- On Thu, 5/11/09, Jeph Herrin wrote:
>> After an estimation, I often want to get a particular
>> element of one of the result matrices, and know I can do
>> it as follows:
>>
>>   reg y x
>>   matrix b = e(b)
>>   di el("b",1,1)
>>
>> But there must be a way to eliminate the middleman b? This
>>
>>   reg y x
>>   di el("e(b)",1,1)
>>
>> doesn't work, but I'm sure there's a way to reference e(b)
>> so that I don't need to store it in another matrix first.
>> In fact, I'm sure I knew it once, even.
> 
> Say the variable whose coefficient you want is called x, 
> than you can refer to it as _b[x] (and its standard error as 
> _se[x]).
> 

*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index