Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


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

Re: st: issue in data analysis


From   "Michael I. Lichter" <[email protected]>
To   [email protected]
Subject   Re: st: issue in data analysis
Date   Fri, 02 Apr 2010 15:40:21 -0400

Ama,

My first suggestion is that you locate a model article or book in your field that does something like what you want to do and use the methods employed in that article/book as a starting point. I'm sure you would also find some data analysis reference books helpful; take a look at those available in the Stata bookstore.

That said, the simplest and least sophisticated approach would be to sum all of the "yes" answers within each of your "axes", creating a summative scale or index of "empowerment". This gives each item equal weight, which might not be appropriate if you think that "access to credit" is a much better indicator of "empowerment" than "increase in revenue". If your variables are coded 0=no, 1=yes, you can create such a scale using -alpha, item generate(varname)-, which will also give you an indication of how well the items cohere, and whether it would be a good idea to drop one or more of the items. The resulting scale is ordinal and can be used as the dependent variable in an ordinal logistic regression (-ologit- or -gologit-) if assumptions are met; often, however, people treat these scales as continuous.

A more sophisticated approach would involve a method like latent class analysis (not available in Stata) or factor analysis (-factor- and -rotate-, not entirely appropriate with dichotomous variables) to examine the structure underlying your measures of "empowerment". You can then create factor scores using -predict- or summative scales using -alpha- to create your outcome variables. You could also use some version of confirmatory factor analysis (not available in Stata) to determine whether your division of "empowerment" into three dimensions is supported by the data. If you search Statalist, I think you will come up with a number of recommended references on factor analysis.

I believe that AbdelRahmen is wrong about -mca- unless your concern is with relationships between the "axes" rather than between these "axes" and the explanatory variables you mentioned.

Michael

amatoallah ouchen wrote:
Thanks a lot Michael for your quick reply,

In fact my survey is about " Measuring  Women's Empowerment through
microfinance, Evidence from Morocco" , in order to do so,  I have
three  principals axes that measure the effect of such program

-Economical empowerment
-welfare empowerment
-the participation if the decision making

of course under each axe there is a yes/no question for example
for the Economical empowerment I have for instance
-Do you have  access to credit? (yes/no)
-is there any significant increase in your revenue?(yes/no)
-etc..

So I want to compute for each individual a dependant variable that
measure this "empowerment" from those yes/no questions , of course
this "empowerment" would be explained by other variables such as: the
level of education, the amount of mortgage, age, etc...

thanks  a lot in advance.
*
*   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/

--
Michael I. Lichter, Ph.D. <[email protected]>
Research Assistant Professor & NRSA Fellow
UB Department of Family Medicine / Primary Care Research Institute
UB Clinical Center, 462 Grider Street, Buffalo, NY 14215
Office: CC 126 / Phone: 716-898-4751 / FAX: 716-898-3536

*
*   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 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index