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

st: RE: looping with roctab

From   "Nick Cox" <>
To   <>
Subject   st: RE: looping with roctab
Date   Fri, 9 Sep 2005 14:58:23 +0100

I would check whether -statsby- can be used 
for this purpose. 

In your code, your call to -egen- is illegal, and 
indeed would be unnecessary if it were legal. 
That is probably where your code fails first. 
(In reporting a problem, it never hurts to say 
precisely where code fails, instead of saying it 
"doesn't seem to do the trick".) 

On the law: The syntax of -egen- always 
includes a call to an -egen- function to 
the right of the assignment sign =. 

On necessity: You are trying to create 
a variable to hold a constant. Outside graphics, 
when this can be a way to solve a problem, 
this is almost never necessary in Stata. 

I would tweak your code in the following 
directions. But, as said, check out -statsby- too. 

use "dataset"

gen year = 1994 + _n in 1/10 
gen arscore = . 

qui forvalues i = 1995/2004 {
	roctab default1 rating if cohort==d(01jan`i'), lorenz summary
	replace arscore = r(gini) if year == `i' 

l year arscore 
summarize arscore


David T. Hamilton
> I have a data set that consists of cohorts of firms, their 
> ratings, and
> their default status (so there are three fields: cohort, default1,
> rating). I am trying to loop over each cohort and take the average of
> the cohort gini scores using roctab rather than pooling the cohort
> data. The problem is that commands that make looping easy are not
> available with roctab (e.g. by, if).
> The code I've written doesn't seem to do the trick. I would appreciate
> any pointers the Stata community could provide. Many thanks.

> Here's my code:
> use "dataset"
> preserve
> mat A=J(10,1,.)
> forvalues i=1/10 {
> egen j=`i'+1994
> keep if cohort==d(01jan`j')
> quietly roctab default1 rating, lorenz summary
> mat A[`i',1]=r(gini)
> restore, preserve
> }
> svmat A, name(arscore)
> summarize arscore

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index