The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Data » Dataset use in SPSS » Prevalence of disability and associated risk factors
Prevalence of disability and associated risk factors [message #10628] Mon, 22 August 2016 04:55 Go to next message
npolle is currently offline  npolle
Messages: 6
Registered: August 2016
Location: MOMBASA, KENYA
Member
I am a beginner in using DHS data. I am interested in using the 2011 Uganda Demographic and Health Survey to determine prevalence of disability and associated risk factors. I have been able to produce similar descriptive statistics to that presented in the final report when using the women and men recode files for example in table 3.1. However when using the Household file (where data on disability is) the statistics i generate are slightly higher (table 2.14). kindly assist.
Secondly can i use the the variable V025 as the strata variable (since V023 and v022 have missing values) when adjusting for complex survey design. This was suggested in one of the forums. Thank you.
Re: Prevalence of disability and associated risk factors [message #10647 is a reply to message #10628] Wed, 24 August 2016 11:45 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3203
Registered: February 2013
Senior Member
Following is a response from Senior DHS Stata Specialist, Tom Pullum:

Are you using Stata? In Stata, the stratum variable can be constructed for this survey with "egen stratum=group(v024 v025)". By itself, v025 is urban/rural residence and cannot be used as the stratum variable. For your other question, please identify specific numbers in tables 3.1 and 2.14 that you believe are inconsistent. Have you tried to match these numbers yourself, using the data files?


Re: Prevalence of disability and associated risk factors [message #10650 is a reply to message #10647] Thu, 25 August 2016 10:33 Go to previous messageGo to next message
npolle is currently offline  npolle
Messages: 6
Registered: August 2016
Location: MOMBASA, KENYA
Member
Thank you Bridgette for your response. Yes I have been able to match the numbers in table 3.1. However I have been unable to match the numbers in table 2.14. Maybe am not doing it correctly. I am using SPSS and here is the syntax I used:
*SELECTING MEMBERS OF AGE >=5.
COMPUTE filter_$=(HV105 >= 5).
VARIABLE LABELS filter_$ 'HV105 >= 5 (FILTER)'.
VALUE LABELS filter_$ 0 'Not Selected' 1 'Selected'.
FORMATS filter_$ (f1.0).
FILTER BY filter_$.
EXECUTE.
*SELECT VALID CASES IN THE VARIABLES SH29 SH25 SH27 SH24 SH26 SH28.
COMPUTE filter_$=(NVALID(SH25) and NVALID(SH29) and NVALID(SH27) and NVALID(SH24) and NVALID(SH26) and NVALID(SH28) ).
* DISTRIBUTION OF HOUSEHOLD MEMBERS OF AGE FIVE YEARS AND OVER BY THE DEGREE OF DIFFICULTY ACCORDING TO FUNCTIONAL AREA.
FREQUENCIES VARIABLES=SH29 SH25 SH27 SH24 SH26 SH28
/ORDER=ANALYSIS.

With this analysis i got a total number of individuals of 36585 instead of 35,226 as shown in table 2.14.
I will also appreciate if am shown how to construct the stratum variable using SPSS.
Thank you.
Nicholas.
Re: Prevalence of disability and associated risk factors [message #10655 is a reply to message #10650] Fri, 26 August 2016 07:22 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3203
Registered: February 2013
Senior Member
Another response from Tom Pullum:

The command in Stata to get the percentage distribution for sh24 in table 2.14, and the total of 35,226 is the following:

use UGPR60FL.dta
tab sh24 [iweight=hv005/1000000] if hv105>=5 & hv105<96 & hv103==1

Here I am excluding cases with hv105>=96, which are basically missing on age. This gives a total of 35,226 weighted cases. Same for sh25 through sh29.

I cannot tell you how to construct a stratum variable in SPSS, because I do not use SPSS. In Stata it would be egen stratum=group(hv024 hv025). Here is another way to do it that does not use egen and group, which are strictly Stata commands.

If you do "tab hv024 hv025" you will see that there are 10 regions and 2 places of residence. There are not 20 combinations, however, because Kampala is completely urban. There are 19. Try this (translated to Stata): "gen stratum=hv024+9*(hv025-1)". There are other ways to do a recode, but this will give the urban parts of the regions as strata 1-10 and the rural parts, excluding Kampala, as 11-19.
Re: Prevalence of disability and associated risk factors [message #10659 is a reply to message #10655] Fri, 26 August 2016 12:04 Go to previous message
npolle is currently offline  npolle
Messages: 6
Registered: August 2016
Location: MOMBASA, KENYA
Member
Thank you so much Bridgette, this has been so helpful.

Nicholas
Previous Topic: Missing data is coded?
Next Topic: Using a sub-sample for analysis
Goto Forum:
  


Current Time: Sun Dec 8 05:34:44 Coordinated Universal Time 2024