Statalist


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

st: about variable conditional substitution


From   Marwan Elkhoury <elkhoury.marwan@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   st: about variable conditional substitution
Date   Wed, 19 Aug 2009 17:42:01 +0200

I hope someone could read and try to find a solution to some simple  
problem, for which I'm unfortunately getting stuck for some reasons,  
though all the tricky parts of the FA analysis were solved.

I was doing some factor analysis for a set of 40 variables, of  
dimension 180, which constitute some macro indicators over 4 different  
groups of countries;
Now after, solving for the factor loadings and estimating the weights  
for the original variables in each group and in each category as well  
as across categories in each group, using matrix algebra to write the  
program, I wanted to link up my results with the given set of countries.

Due to matrix manipulations, my final country variable is no more  
related to the original one, in the sense that when I compare the  
final results with respect to the different group of countries, it  
does not relate to them anymore; in other words, the initial country  
variable and the final country variable, which is a plain substitution  
of one for the other, does not express anymore the one and same  
variable and they act as though they were completely independent.

Though I tried to use :
replace final_country = original_countryinal_country]
and some other tricks, it does the substitution correctly but it does  
not link the value of the country name to the other.

I need to tell the program that the final country variable is the  
initial country variable by substituting, element by element, the  
final one for the original one, discarding the missing elements, given  
that the initial country variable is just an 'encoding' of the string  
country name variable.

If anyone gets my message and has some way of knowing how to "re-link"  
these two variables together, I would greatly appreciate any hint or  
indication.

regards,


*
*   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