| Re: Contraceptive calendar in Nigeria 2013 dataset [message #6841 is a reply to message #6788] |
Tue, 21 July 2015 09:41   |
Liz-DHS
Messages: 1516
Registered: February 2013
|
Senior Member |
|
|
Dear User,
Here is a response from one of our data processing staff, Mianmian Yu. Her responses are in bold.
1. In the recode manual, it states that the calendar data are stored as single variables of 80 characters. In the Nigeria 2013 dataset, it appears that there are only 67 characters. Is this correct? '80 characters' means the maximum length of the calendar variables. The factual number of characters that appear in the variables depends on the month and year of interview.
2. Our next question refers to vcal_1. It is my understanding that the earliest positions in the string variable relate to the most recent times and the end of the string relates to further back in time. When we use the code below to destring the variable we find that the last positions for some respondents are blank. It is my understanding that the blank positions correspond to the months after the month of the interview, thus shouldn't the blanks appear in the first positions because they would be the most recent times? Spaces for the months after the interview month must be blank because the survey was finished before then. The first non-blank position should represent the most recent status of contraceptive use.
3. Our next question refers to vcal_2. When we use the code below to destring vcal_2, we find that when a respondent provides a reason for contraceptive discontinuation it appears in the 1st position regardless of when the respondent actually discontinued the contraceptive method. I would expect that the reason for discontinuation would be entered in the position that corresponds to the date of discontinuation from vcal_1. Is our code below used to destring vcal_2 incorrect? The reason for discontinuation is recorded at the month where a contraceptive method is switched off. If there appears to be an exceptional case, please let us know its ID ( cluster #, HH# and line # ) and we will be happy to investigate it.
4. Our last question refers to vcal_3, which according to the recode manual is the column that records episodes of marriage. Does the Nigeria 2013 dataset include vcal_3? If so, is the variable called something else? We can't seem to find it in the dataset. Marriage status was not asked in the calendar of Nigeria DHS 2013.
If you have additional questions, please feel free to post again.
Thank you!
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
