| NFHS III [message #17603] |
Thu, 25 April 2019 12:31  |
dgodha
Messages: 44
Registered: November 2016
Location: India
|
Member |
|
|
Hello,
I am working on calendar data using the latest version - IAIR52.dta. Strangely enough, in this version, v017 has no values but shows two categories on tabulation. In the earlier version- IAIR50.dta, v017 had 2 values: 1201 and 1213. In the latest version, using v008 and v019, I get the same 2 values for the start of the calendar. But, a quick look at vcal_1 shows no blanks towards the right end.
The tutroial on calendar data explicitly says that the start of a calendar is constant for all cases (row 80) while the end of the calendar may vary based on the month of interview.
Am I correct in my understanding that the calendar data for India DHS 2005-06 followed two start dates and as such needs to be realigned if I want all cases to correspond to the same calendar month?
I will appreciate your response (even a simple yes/no will be of help) and thank you in advance.
Deepali
Deepali
|
|
|
|
| Re: NFHS III [message #17660 is a reply to message #17603] |
Fri, 03 May 2019 03:49  |
dgodha
Messages: 44
Registered: November 2016
Location: India
|
Member |
|
|
Hello,
My calculations below show the above is possibly true. I will still appreciate if someone can confirm it. The confusion is that all observations end in row 80. In other words, the start of calendar is row 80 for all observations despite being different. This means that I will have to realign the calendar to get all observations to correspond to any particular month or period.
Thanks
Deepali
[Updated on: Fri, 03 May 2019 03:50] Report message to a moderator |
|
|
|
| Re: NFHS III [message #17713 is a reply to message #17660] |
Thu, 09 May 2019 13:44 |
Trevor-DHS
Messages: 803
Registered: January 2013
|
Senior Member |
|
|
I will try to answer each of the questions:
1) Yes, v017 has two values. Unfortunately these don't show when you just tab v017 as there list a label that has blank entries for v017. Use to drop the label and then see the values.
2) In almost all surveys (with the exception of this and other India surveys) the data collection is done with a fixed start point for the calendar. For the India 2005-06 survey, the data collection was done in two phases, and the start point of the calendar varies by 1 year between the two phases.
3) If you want the two calendars to line up for exactly the same months, then yes, you do need to realign the data from one phase. However, for the majority of analyses we are using time periods relative to the date of interview, and in that case it is typically not necessary to realign the data. Your use may be different and you may need to realign the data. This can be done simply by creating a new variable and adding 12 blanks to the front of the calendar string for the interviews in phase 1.
I hope this helps.
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
