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: Analysis on subset of Demographic and Health Survey (DHS) data
From
Friedrich Huebler <[email protected]>
To
[email protected]
Subject
Re: st: Analysis on subset of Demographic and Health Survey (DHS) data
Date
Sun, 27 Oct 2013 23:09:33 -0400
French,
All DHS datasets should contain the same information, independent of
the file format (flat, hierarchical, etc.). Quote from the MEASURE DHS
website: "Most researchers use the flat file designed for the
statistical software that they intend to use for analysis."
https://www.measuredhs.com/data/File-Formats.cfm
Friedrich
On Sun, Oct 27, 2013 at 2:00 PM, french Smith <[email protected]> wrote:
> Dear Friedrich,
>
> Apologies for not having noted my Stata version, and for other
> omissions of important info.
>
> My Stata output is below.
>
> I use version 10.
>
> I expected 8,081 cases because page 14 of the report says there were
> 8,081 children aged 0-4.
>
> Thank you for reminding me that Stata is case sensitive - I'm happy
> that there is an easy fix to that issue.
>
> When you analyse DHS, do you analyse the same file type I was using? I
> saw that in one of the posts you note you prefer using the flat file.
>
> Best,
>
> French
>
>
> . set mem 450m
> (460800k)
>
> .
> . set matsize 800
>
> .
> . use TZKR62DT/TZKR62FL.DTA
>
> . *v022 is sample stratum number
>
> . gen stratid = v022
>
> .
> . *v021 is PSU
>
> . gen psu = v021
>
> .
> . *v005 is sample weight
>
> . gen weight = v005/1000000
>
> .
> . svyset psu [pw=weight], strata(stratid)
>
> pweight: weight
> VCE: linearized
> Single unit: missing
> Strata 1: stratid
> SU 1: psu
> FPC 1: <zero>
>
> . *Just to compare totals. Should get 8,081 for Tanzania DHS 2010 not 8,023!
>
> . tab stratid
>
> stratid | Freq. Percent Cum.
> ------------+-----------------------------------
> 1 | 30 0.37 0.37
> 2 | 83 1.03 1.41
> 3 | 38 0.47 1.88
> 4 | 41 0.51 2.39
> 5 | 63 0.79 3.18
> 6 | 58 0.72 3.90
> 7 | 194 2.42 6.32
> 8 | 27 0.34 6.66
> 9 | 45 0.56 7.22
> 10 | 39 0.49 7.70
> 11 | 40 0.50 8.20
> 12 | 46 0.57 8.77
> 13 | 57 0.71 9.49
> 14 | 39 0.49 9.97
> 15 | 44 0.55 10.52
> 16 | 40 0.50 11.02
> 17 | 36 0.45 11.47
> 18 | 16 0.20 11.67
> 19 | 83 1.03 12.70
> 20 | 54 0.67 13.37
> 21 | 27 0.34 13.71
> 23 | 15 0.19 13.90
> 24 | 278 3.47 17.36
> 25 | 34 0.42 17.79
> 26 | 84 1.05 18.83
> 27 | 266 3.32 22.15
> 28 | 176 2.19 24.34
> 29 | 112 1.40 25.74
> 30 | 181 2.26 27.99
> 31 | 188 2.34 30.34
> 32 | 192 2.39 32.73
> 33 | 13 0.16 32.89
> 34 | 200 2.49 35.39
> 35 | 154 1.92 37.31
> 36 | 213 2.65 39.96
> 37 | 189 2.36 42.32
> 38 | 233 2.90 45.22
> 39 | 312 3.89 49.11
> 40 | 445 5.55 54.66
> 41 | 310 3.86 58.52
> 42 | 292 3.64 62.16
> 43 | 488 6.08 68.24
> 44 | 300 3.74 71.98
> 45 | 363 4.52 76.51
> 46 | 398 4.96 81.47
> 47 | 264 3.29 84.76
> 48 | 327 4.08 88.83
> 49 | 239 2.98 91.81
> 50 | 56 0.70 92.51
> 51 | 331 4.13 96.63
> 52 | 270 3.37 100.00
> ------------+-----------------------------------
> Total | 8,023 100.00
>
> .
> . *keep if complete response
>
> . keep if v015==1
> (0 observations deleted)
>
> . *keep if child lives with respondent
>
> . keep if b9==0
> (824 observations deleted)
>
>
> On Sun, Oct 27, 2013 at 9:48 AM, Friedrich Huebler <[email protected]> wrote:
>> French,
>>
>> I have to retract my comment about the filename. The -use- command is
>> valid if you have a folder TZKR62DT that contains the file
>> TZKR62FL.DTA.
>>
>> Sorry for my mistake,
>>
>> Friedrich
>>
>> On Sun, Oct 27, 2013 at 9:22 AM, Friedrich Huebler <[email protected]> wrote:
>>> French,
>>>
>>> Please have a look at the Statalist FAQ (link at the bottom of this
>>> message). Some excerpts:
>>>
>>> "The current version of Stata is 13.0. Please specify if you are using
>>> an earlier version"
>>>
>>> The fact that you use -set mem- indicates that you have Stata 11 or older.
>>>
>>> "Say exactly what you typed and exactly what Stata typed (or did) in response."
>>>
>>> I assume that you did not type "use TZKR62DT/TZKR62FL.DTA" because
>>> that is not a valid filename in most operating systems. You also don't
>>> explain why the command "keep if B9==0" is not recognized because you
>>> omitted the output from Stata.
>>>
>>> Now to your questions. The data for children under 5 from the Tanzania
>>> DHS 2010 has only 8023 observations. It is not clear why you expect
>>> 8081 observations. The survey report, which you cite as reference, has
>>> 480 pages.
>>>
>>> The command -keep if B9==0- will yield an error message because the
>>> variable B9 doesn't exist in the data. Stata is case-sensitive and the
>>> correct variable name is b9.
>>>
>>> Friedrich
>>>
>>> On Sat, Oct 26, 2013 at 12:53 PM, french Smith <[email protected]> wrote:
>>>> Dear STATA crowd,
>>>>
>>>> I wish to analyze the Demographic and Health Surveys (DHS) data on
>>>> children under five.
>>>>
>>>> I have set up my analysis as follows, using at Tanzania DHS 2010 as a
>>>> starting point. But I have two questions:
>>>>
>>>> 1. Why do I have 8,023 cases not the 8,081 that the report (the report
>>>> and dataset are available at
>>>> http://www.measuredhs.com/what-we-do/survey/survey-display-345.cfm)
>>>> says exit?
>>>>
>>>> 2. How do I restrict the analysis to V465, which was a question asked
>>>> only of respondents with a youngest child under five living with them?
>>>>
>>>> Maybe questions, sorry, but my STATA mind needs jogging!
>>>>
>>>> Thanks!
>>>>
>>>> French
>>>>
>>>>
>>>> set mem 450m
>>>>
>>>> set matsize 800
>>>>
>>>> use TZKR62DT/TZKR62FL.DTA
>>>>
>>>> *v022 is sample stratum number
>>>> gen stratid = v022
>>>>
>>>> *v021 is PSU
>>>> gen psu = v021
>>>>
>>>> *v005 is sample weight
>>>> gen weight = v005/1000000
>>>>
>>>> svyset psu [pw=weight], strata(stratid)
>>>>
>>>> *Just to compare totals. Should get 8,081 for Tanzania DHS 2010 not 8,023!
>>>> tab stratid
>>>>
>>>> *keep if complete response
>>>> keep if v015==1
>>>>
>>>> *keep if child lives with respondent. But this command isn’t recognized!
>>>> keep if B9==0
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/