The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Countries » India » urban-rural variable
urban-rural variable [message #4226] Tue, 21 April 2015 10:14 Go to next message
menonidhi is currently offline  menonidhi
Messages: 1
Registered: April 2015
Location: India
Member
In my study, I have two variables - V025: which classifies place of residence as Urban/Rural and S025: which classifies residence as Mega city/Large City/Small City/Large Town/Small town/Rural.

I was able to extract the former only for females. To check if I could could substitute V025 with S025( which I was able to extract for both the sexes), I checked the distribution of urban and rural population among females. I clubbed the values of Mega city/Large City/Small City/Large Town/Small town as Urban.

I however observe few discrepancies. The number of females in the urban area is 18,442 and rural is 20,815. In the second variable, the number of females in rural area is 23,529. I have summed the values of mega city,large city,small city, large town and small town which comes to 15,728.
I observe an exact difference of 2,714 i.e. (23,529-20,815=2714) and (18,442-15,728 =2,714). I have attached my results for any reference for the values mentioned above.

Is there a reason for this difference.
Re: urban-rural variable [message #4243 is a reply to message #4226] Fri, 24 April 2015 15:21 Go to previous message
Trevor-DHS is currently offline  Trevor-DHS
Messages: 787
Registered: January 2013
Senior Member
It wasn't very clear from your explanation which survey and datasets you are using, but it appears that you are using the HIV tests results data, matched with other datasets. However it appears that there is a problem with the matching of your data. I would expect the number of women and men to be roughly equal, but you have twice as many men as women. Below is a short piece of code for matching with the persons recode file, and for checking the sex distribution:
cd "C:\Data\DHS_Stata"
use "IAPR52FL.dta"
clonevar hivclust = hv001
clonevar hivnumb = hv002
clonevar hivline = hvidx
merge 1:1 hivclust hivnumb hivline using "IAAR51FL.dta"
keep if _merge==3
tab hv104
tab hv025
tab sh025
tab sh025 hv025

From this, I received the following output:
. tab hv104

     sex of |
  household |
     member |      Freq.     Percent        Cum.
------------+-----------------------------------
       male |     51,641       48.88       48.88
     female |     54,016       51.12      100.00
------------+-----------------------------------
      Total |    105,657      100.00

. tab hv025

    type of |
   place of |
  residence |      Freq.     Percent        Cum.
------------+-----------------------------------
      urban |     52,697       49.88       49.88
      rural |     52,960       50.12      100.00
------------+-----------------------------------
      Total |    105,657      100.00

. tab sh025

city\town\c |
 ountryside |      Freq.     Percent        Cum.
------------+-----------------------------------
  mega city |      8,696        8.23        8.23
 large city |     21,989       20.81       29.04
 small city |      6,352        6.01       35.05
 large town |      1,671        1.58       36.64
 small town |     13,989       13.24       49.88
      rural |     52,960       50.12      100.00
------------+-----------------------------------
      Total |    105,657      100.00

. tab sh025 hv025

city\town\ |   type of place of
countrysid |       residence
         e |     urban      rural |     Total
-----------+----------------------+----------
 mega city |     8,696          0 |     8,696 
large city |    21,989          0 |    21,989 
small city |     6,352          0 |     6,352 
large town |     1,671          0 |     1,671 
small town |    13,989          0 |    13,989 
     rural |         0     52,960 |    52,960 
-----------+----------------------+----------
     Total |    52,697     52,960 |   105,657 
As you can see in my tables (unweighted) there are about 49% men and 51% women, which is roughly what I would expect.
Also, as you can see the recoding of urban and rural makes sense in the last table.
Previous Topic: Dowry in NFHS
Next Topic: Village identifiers in NFHS3
Goto Forum:
  


Current Time: Fri Mar 29 02:38:26 Coordinated Universal Time 2024