| Additional information on Cambodia DHS 2014 report and data set [message #15388] |
Tue, 10 July 2018 07:52  |
tasnuva
Messages: 6
Registered: July 2018
|
Member |
|
|
I have received DHS 2014 Cambodia data set from DHS program to analyze and report on early child development index in Cambodia. I need a clarification about the age in month of children variable. According to published DHS 2014 Cambodia report, the number of children aged 36-59 months included for analysis on early child development index (ECDI) was 2617. By contrast, I got responses on ECDI related variables from 3545-3551 children in the data set that I received from DHS program. I have recoded a new variable on age of children having 36-59 month age using 'age in month' variables (W1$1 HW1$2 HW1$3 HW1$4 HW1$5 HW1$6). This new variable gives a sample of 1665 children. I am attaching here an SPSS output file. I have read the sample weighting report in the appendix of DHS 2014 report, but do not find information on child development topic.
Would you please, mind to let me know the variable name on 'age in month of children' variable that can be used to analyze ECDI variables from Cambodia data set.
Please, feel free to ask or write (email:tasnuva.wahed@gmail.com) me if needs.
Looking forward to your kind reply.
Dr. Tasnuva Wahed
|
|
|
|
|
|
| Re: 'age in month of children' variable on child development in Cambodia DHS 2014 report and data set [message #15423 is a reply to message #15389] |
Tue, 17 July 2018 12:04  |
Trevor-DHS
Messages: 805
Registered: January 2013
|
Senior Member |
|
|
There are a number of issues:
1) Data were collected for children age 3-5 years (36-71 months), but are presented for children age 36-59 months, so you need tyo select only those cases.
2) You are looking at the HW1 series of variables for age in months, but this is only for children weighed and measured, and I believe that was just a subsample of the full sample, so you won't find age for all children there.
3) You are setting your variable age_36_59 to 1 if any of the children in the HW1 series is age 36-59. Rather you need to identify the correct child. If you look at the questionnaire at the end of the report you will see in filter 570 that based on Q217 the youngest child age 3, 4 or 5 years was selected.
Below I am providing some code for correctly setting the age of the child and selecting the cases used in the tabulation:
* Open the dataset.
get file='KHIR73FL.SAV'.
* Age of youngest child age 3-5.
* Create a vector of the date of birth variables from the birth history.
vector B3 = B3$01 to B3$20.
* Select the birth date of the index child, based on S570A. If 1 uses B3$01, if 2 uses B3$02, etc.
if (not sysmis(S570A)) birthdate = B3(S570A).
* calculate the age of the child.
compute age = v008 - birthdate.
compute agegroup = trunc(age/12).
variable labels agegroup 'Age in months'.
value labels agegroup 3 '36-47' 4 '48-59' 5 '60-71'.
* filter for children age 36-59 months.
use all.
compute filter_$=(age < 60).
filter by filter_$.
* weight the data.
compute wgt = v005/1000000.
weight by wgt.
* check the counts.
freq agegroup. |
|
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
