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

st: RE: Tables and row/column percentages

From   "Nick Cox" <[email protected]>
To   <[email protected]>
Subject   st: RE: Tables and row/column percentages
Date   Tue, 28 Oct 2003 19:02:57 -0000

The short answer is that -table- does not 
provide row and column percents. 
I guess that the reason for that, historically 
and also logically, is that -tabulate- 
was already in existence as a command 
able to do that, at least for one- and two-way tables, 
although -tabulate- does not accept pweights. 
However, some experiments suggest that you 
get the same results using iweights.  

Various hacks exist to tackle part of this lack 
directly. -tablecol- from Nick Winter 
and -tablepc- from myself may be found on 
SSC. Nick's approach is a wrapper approach; 
mine is a prepare-variable-in-advance approach. 

There is a general discussion of tables 
in Stata in "Problems with tables, Part I"
Stata Journal. Stata Journal 3(3):309--324 (2003).      
from which it is apparent that "Problems with 
tables, Part II" will follow. I understand 
that there are no current plans for "Problems 
with tables, Part III". 

Following on the runaway success of "The matrix", 
"The matrix transposed", "The matrix eigenanalysed", 
etc., other media may be used to publish this material. 

I don't understand how using -table- would be 
more efficient than using -tabulate-. 

[email protected] 

> -----Original Message-----
> I have a question concerning the "table" command, and 
> actually, the usage of
> tables in general in STATA.  I am workign with a dataset in 
> which I would
> like to examine variables such as marriage by a number of different
> indicator variables, such as district, village type (i.e. 
> intervention vs.
> control), age, etc.   Also, I need to examine the data 
> using a pweight.
> I was wondering if there is anyway to get row/column 
> percentages using the
> "table"" command.  I know that I could get these 
> percentages using the "tab"
> command, but I find it inefficient to run lots of tabs and 
> have the output
> be scattered all over the place.
> In sum, I was wondering if there is anyway to get 
> row/column percentages for
> a number of different variables in one step.  

*   For searches and help try:

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