The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Data » General Data Questions » 2013 Namibia DHS data: Numbers not matching
2013 Namibia DHS data: Numbers not matching [message #13540] Fri, 17 November 2017 07:42 Go to next message
jwichmannza is currently offline  jwichmannza
Messages: 3
Registered: October 2017
Location: South Africa
Member
I am using the 2013 Namibia DHS data in one of the projects of my MPH students here at the University of Pretoria.

Please refer to Table 11.5 in the report.

I do not get these numbers at all when I use the children's recode data set, namely NMKR61FL.


e.g. below are the numbers/frequencies for the age categories I get


The SAS System 09:45 Thursday, November 16, 2017 16

The FREQ Procedure

Cumulative Cumulative
agecat Frequency Percent Frequency Percent
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
1 271 13.93 271 13.93
2 251 12.90 522 26.84
3 430 22.11 952 48.95
4 374 19.23 1326 68.17
5 323 16.61 1649 84.78
6 296 15.22 1945 100.00

Frequency Missing = 3101



data xxx ;
set xxx ;

agecat = . ;

if hw1 ne . and hw1 > 47 then agecat = 6;
if hw1 ne . and hw1 < 48 then agecat = 5;
if hw1 ne . and hw1 < 36 then agecat = 4;
if hw1 ne . and hw1 < 24 then agecat = 3;
if hw1 ne . and hw1 < 12 then agecat = 2;
if hw1 ne . and hw1 < 6 then agecat = 1;

run;




I get many missing values for the age categories, even in the FREQ file it is the same.
Re: 2013 Namibia DHS data: Numbers not matching [message #13554 is a reply to message #13540] Fri, 17 November 2017 13:59 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3017
Registered: February 2013
Senior Member


Following is a response from Senior DHS Stata Specialist, Tom Pullum:

I believe there are two possible reasons why you are not getting a match. The first is that you need to use sampling weights. The weights are given by v005. Please read up on weights elsewhere in the forum or in our documentation.

Second, the KR file has all births in the past five years, including children who have died. To get the surviving children you need to include (in Stata, not SPSS) "if b5==1". I believe you will be able to replicate the table with those changes.

Please let us know if you still cannot match.

Re: 2013 Namibia DHS data: Numbers not matching [message #13564 is a reply to message #13554] Mon, 20 November 2017 03:34 Go to previous messageGo to next message
jwichmannza is currently offline  jwichmannza
Messages: 3
Registered: October 2017
Location: South Africa
Member
Hi again

See the FRQ file (uploaded here)

Item HW1: Child's age in months
... tbd-name: '.NMKR61FL_DICT.RECORD1.HW1'
_____________________________ _____________ _____________
Categories Frequency CumFreq % Cum % Net %|cNet %
_______________________________ _____________________________ _____________ _____________
0 18 18 0.4 0.4 0.9 0.9
1 55 73 1.1 1.4 2.8 3.8
2 52 125 1.0 2.5 2.7 6.4
3 43 168 0.9 3.3 2.2 8.6
4 58 226 1.1 4.5 3.0 11.6
5 45 271 0.9 5.4 2.3 13.9
6 46 317 0.9 6.3 2.4 16.3
7 44 361 0.9 7.2 2.3 18.6
8 44 405 0.9 8.0 2.3 20.8
9 40 445 0.8 8.8 2.1 22.9
10 44 489 0.9 9.7 2.3 25.1
11 33 522 0.7 10.3 1.7 26.8
12 35 557 0.7 11.0 1.8 28.6
13 52 609 1.0 12.1 2.7 31.3
14 34 643 0.7 12.7 1.7 33.1
15 32 675 0.6 13.4 1.6 34.7
16 44 719 0.9 14.2 2.3 37.0
17 42 761 0.8 15.1 2.2 39.1
18 31 792 0.6 15.7 1.6 40.7
19 28 820 0.6 16.3 1.4 42.2
20 32 852 0.6 16.9 1.6 43.8
21 40 892 0.8 17.7 2.1 45.9
22 25 917 0.5 18.2 1.3 47.1
23 35 952 0.7 18.9 1.8 48.9
24 32 984 0.6 19.5 1.6 50.6
25 41 1025 0.8 20.3 2.1 52.7
26 25 1050 0.5 20.8 1.3 54.0
27 33 1083 0.7 21.5 1.7 55.7
28 36 1119 0.7 22.2 1.9 57.5
29 34 1153 0.7 22.8 1.7 59.3
30 28 1181 0.6 23.4 1.4 60.7
31 37 1218 0.7 24.1 1.9 62.6
32 33 1251 0.7 24.8 1.7 64.3
33 29 1280 0.6 25.4 1.5 65.8
34 30 1310 0.6 26.0 1.5 67.4
35 16 1326 0.3 26.3 0.8 68.2
36 29 1355 0.6 26.9 1.5 69.7
37 30 1385 0.6 27.4 1.5 71.2
38 39 1424 0.8 28.2 2.0 73.2
39 28 1452 0.6 28.8 1.4 74.7
40 23 1475 0.5 29.2 1.2 75.8
41 26 1501 0.5 29.7 1.3 77.2
42 24 1525 0.5 30.2 1.2 78.4
43 25 1550 0.5 30.7 1.3 79.7
44 26 1576 0.5 31.2 1.3 81.0
45 14 1590 0.3 31.5 0.7 81.7
46 34 1624 0.7 32.2 1.7 83.5
47 25 1649 0.5 32.7 1.3 84.8
48 30 1679 0.6 33.3 1.5 86.3
49 26 1705 0.5 33.8 1.3 87.7
50 19 1724 0.4 34.2 1.0 88.6
51 32 1756 0.6 34.8 1.6 90.3
52 32 1788 0.6 35.4 1.6 91.9
53 22 1810 0.4 35.9 1.1 93.1
54 24 1834 0.5 36.3 1.2 94.3
55 26 1860 0.5 36.9 1.3 95.6
56 18 1878 0.4 37.2 0.9 96.6
57 20 1898 0.4 37.6 1.0 97.6
58 21 1919 0.4 38.0 1.1 98.7
59 26 1945 0.5 38.5 1.3 100.0
_______________________________ _____________________________ _____________ _____________
NotAppl 3101 5046 61.5 100.0
_______________________________ _____________________________ _____________
TOTAL 5046 5046 100.0 100.0




So 3101 missing values for the child's age (in month) variable.

Only 228 children were reported not to be alive anymore, see below, again from FRQ file.

So why so many missing values for the HW1 variable then?


____________________________________________________________ ____________________
Item B5: Child is alive
... tbd-name: '.NMKR61FL_DICT.RECORD1.B5'
_____________________________ _____________
Categories Frequency CumFreq % Cum %
_______________________________ _____________________________ _____________
0 No 228 228 4.5 4.5
1 Yes 4818 5046 95.5 100.0
_______________________________ _____________________________ _____________
TOTAL 5046 5046 100.0 100.0

  • Attachment: NMKR61FL.FRQ
    (Size: 1.70MB, Downloaded 412 times)
Re: 2013 Namibia DHS data: Numbers not matching [message #13695 is a reply to message #13564] Fri, 08 December 2017 11:10 Go to previous messageGo to next message
Trevor-DHS is currently offline  Trevor-DHS
Messages: 787
Registered: January 2013
Senior Member
The HW series of variables are for children measured for height and weight. In this survey height and weight was measured only for a subsample of children, and thus the age in this series of variables is provided only for this subsample.

The recommended method of calculating age in months for children is to calculate it using v008 and b3 as follows:

age = v008 - b3

See the note about calculating age of children on our website, which also describes recent changes we have made in the calculation of age in the most recent surveys.
Re: 2013 Namibia DHS data: Numbers not matching [message #13705 is a reply to message #13695] Mon, 11 December 2017 12:28 Go to previous message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3017
Registered: February 2013
Senior Member
Following is another response (from Senior DHS Stata Specialist, Tom Pullum):

In the Namibia 2013 survey, it appears that there was a household-level selection criterion given by hv042, "household selected for hemoglobin", that refers to much more data than just the hemoglobin measurements. The information about children under five is only included for households with hv042=1. The following lines can be used to reduce the KR file to those households.

These households were randomly selected, so estimates from the KR file should be unbiased. You will find more information on this in the main report on this survey.

set more off
use e:\DHS\DHS_data\PR_files\NMPR61FL.dta, clear
keep if hvidx==1
keep hv001 hv002 hv042
tab hv044
rename hv001 v001
rename hv002 v002
sort v001 v002
save e:\DHS\DHS_data\scratch\temp.dta, replace

use e:\DHS\DHS_data\KR_files\NMKR61FL.dta, clear
keep v001 v002 hw1 b*
sort v001 v002
merge v001 v002 using e:\DHS\DHS_data\scratch\temp.dta
tab _merge
keep if _merge==3
tab hw1 hv042,m
tab hw1 b5 if hv042==1,m

[Updated on: Mon, 11 December 2017 13:39]

Report message to a moderator

Previous Topic: Is the DHS data a longitudinal one?
Next Topic: Question on Benin's Raw Service Availability Data 2006
Goto Forum:
  


Current Time: Thu Mar 28 16:11:05 Coordinated Universal Time 2024