| Re: Replicating Table 9.3 in EDHS2011 [message #11915 is a reply to message #11911] |
Thu, 02 March 2017 09:58  |
Bridgette-DHS
Messages: 3230
Registered: February 2013
|
Senior Member |
|
|
Following is a response from DHS Senior Research Associate, Lindsay Mallick:
The denominator "Women with a live birth in the last 5 years who received ANC for their most recent pregnancy" includes women who responded that they didn't know how many ANC visits they had. We recommend that you also look at the videos recently posted on troubleshooting/matching DHS numbers in the final reports.
gen wt = v005/1000000
*generate denominator 1 for table 9.3: any woman who had a birth in the last 5 years
recode v208 (0=0) (1/10=1) (else = .), gen(birth5)
label var birth5 "Had a birth in the last five years"
*generate denominator 2 for table 9.3: any woman who had a birth in the last 5 years who received anc for most recent birth
tab m14_1, m
recode m14_1 (0=0 "no anc") (1/20 = 1 "anyanc") (else = 9 "dk/miss"), gen(anyanc)
replace anyanc = . if birth5==0
*check to see if coding was done correctly
tab m14_1 anyanc [iw=wt], m
**anyanc (excluding don't know and missing), weighted = 3365
**denominator for women with ANC in final report = 3391
**the difference is the 27 women with "don't know responses" to question on number of visits
drop anyanc
**recode to match final report
recode m14_1 (0=0 "no anc") (1/98 = 1 "at least 1 visit or don't know how many") (else = .), gen(anyanc)
replace anyanc2 = . if birth5==0
*check for errors
tab m14_1 anyanc, m
*check weighted number against final report
tab anyanc [iw=wt] //matches!
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
