Equity index [message #10121] |
Wed, 29 June 2016 11:02 |
DHS user
Messages: 111 Registered: February 2013
|
Senior Member |
|
|
I would like to create an equity index score using education, wealth and residence as proxy measure. I am using pooled data from the 2003 and 2014 Kenya DHS file for women.
Would you please advise how to go about this in STATA or direct me to the right path?
|
|
|
Re: Equity index [message #10122 is a reply to message #10121] |
Wed, 29 June 2016 11:04 |
Bridgette-DHS
Messages: 3185 Registered: February 2013
|
Senior Member |
|
|
Following is a response from Senior DHS Stata Specialist, Tom Pullum:
First, I would keep education, wealth, and residence separate, rather than trying to combine them in some way. (Maybe I misunderstood, but you seemed to be thinking of combining them.) Sub-national region is also usually important. You could combine them but they have their own separate interpretations.
Second, possible measures are the odds ratios of the highest vs. lowest categories (for place of residence there are only two categories, of course), or the difference between the highest and lowest categories, or a Gini coefficient for the distribution across all categories (if more than two). Stata has a built-in version of the Gini coefficient.
You can also interpret the multiple R-squared and the proportion of deviance explained as measures of equity (low values mean high equity).
Here at DHS we are working on a report in equity and changes in equity of all the MCH indicators in a large number of surveys. That will be finished by September. There are other recent non-DHS publications on the topic. Tell us if you have difficulty finding them.
|
|
|