| Decile Share Thru Wealth Index Factor (stata) [message #2351] |
Sat, 07 June 2014 09:10  |
smileyc
Messages: 14
Registered: June 2014
|
Member |
|
|
Hi,
I would like to seek help in generating wealth index by deciles.
I am using xtile decile [weight=weight]=hv271, nq(10) but I think this is not the write syntax in stata
Hope to hear from you.
Thanks!
|
|
|
|
|
|
|
|
|
|
| Re: Decile Share Thru Wealth Index Factor (stata) [message #2357 is a reply to message #2355] |
Sat, 07 June 2014 22:57  |
smileyc
Messages: 14
Registered: June 2014
|
Member |
|
|
Hello,
Thank you for taking time to help me. I am trying to get the share of the richest 10% of the schooling years.
I first tried to compare the wealth index (hv270) with the variable I created "xtile quint [weight=weight] = hv271, nq(5)"
to check if I can generate decile share using hv271 but the results of hv270 is not similar with the results of the quint variable.
Thanks again.
[Updated on: Sat, 07 June 2014 23:00] Report message to a moderator |
|
|
|
| Re: Decile Share Thru Wealth Index Factor (stata) [message #2358 is a reply to message #2357] |
Sun, 08 June 2014 16:09 |
Reduced-For(u)m
Messages: 292
Registered: March 2013
|
Senior Member |
|
|
Hmmmm... so looking through the 2011 Bangladesh report (I don't know which country you are looking at, but it is likely similar), I see this:
Once the index is computed, national-level wealth quintiles (from lowest to highest) are obtained by
assigning the household score to each de jure household member, ranking each person in the population by
his or her score, and then dividing the ranking into five equal categories, each comprising 20 percent of the
population.
Which indicates to me that perhaps you need to implicitly weight by household size to back out the proper wealth quintiles (because each household member gets the household score, and then all people surveyed are ordered to find the quintiles instead of each household surveyed). So you might try using the household member recode to do this breakdown - that would also increase the probability you are using the same weights as the DHS people - I don't know for sure, but it sounds like they do this using the individual household member list, and not the household schedule (which is where the HV variables come from).
Let me know if that works.
|
|
|
|
|
|
| Re: Decile Share Thru Wealth Index Factor (stata) [message #2368 is a reply to message #2351] |
Mon, 09 June 2014 14:15 |
Liz-DHS
Messages: 1516
Registered: February 2013
|
Senior Member |
|
|
Dear User,
Here also is a response from Dr. Shea Rutstein. Dr. Rutstein is not a STATA expert, but he offered the following:
I am not the expert for STATA, but I think the code should be:
Use xtile pctile decile=hv271 [weight= hv012*hv005/1000000], nq(10). The quantiles (percentiles) are based on the dejure household population so the weight is the household weight multiplied by the number of dejure household members. I don't think that this procedure gives the cutpoints but rather recodes hv271 into 10 categories. Use pctile to get the cutpoints.
|
|
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
