The DHS Program User Forum: Fertility » Not matching with Table

Home » Topics » Fertility » Not matching with Table

Show: Today's Messages :: Show Polls :: Message Navigator

Not matching with Table [message #28208]

Wed, 29 November 2023 07:01

Rupon
Messages: 16
Registered: October 2023

Member

Hi,

I am using NFHS-5 (2019-21) for India. I tried calculating the number of MOST RECENT LIVE BIRTH (midx==1) not delivered in health facility (TABLE 8.14 in report). Following a reply to another comment at https:// userforum.dhsprogram.com/index.php?t=msg&goto=18260& srch=institutional%20delivery&SQ=c6f7e7a62ac4917bd6ea269 bff619101&, I modified and ran the following command in Stata:

*generating weight
gen wt = v005/1000000

*recoding delivery place
recode m15 (11 = 4 "Own home") (12 = 5 "Other home") (13 = 6 "Parents' home") ///
(21/27 = 1 "Public Sector") (33 = 2 "NGO") (31/32 = 3 "Private") ///
(96 . =7 "Other"), gen(facdel)

*recoding delivery place
recode facdel (4 5 6 = 2 "Home") (1 2 3 = 1 "Facility") (7 = 3 "Other"), gen(facdel2)

*institutional and non-institutional
recode facdel2 (1 = 1 "Institutional") (2 3 =2 "Non-institutional"), gen (facdel3)

*tabulation for most recent live birth
tab facdel3 [iw=wt] if midx==1

*RESULT: Total NON-INSTITUTIONAL BIRTHS did not match with the table [in Table 17016, vs. our result 17395]. It seems that 'OTHER' is taken as 'INSTITUTIONAL' in calculating the report result.

Kindly note that for all live births, the calculated result matches with the report Table 8.13 with this command, however.

Your support is highly appreciated.

Thank you

Regards
Rupon Basumatary

Report message to a moderator

Re: Not matching with Table [message #28268 is a reply to message #28208]

Wed, 06 December 2023 16:56

Janet-DHS
Messages: 891
Registered: April 2022

Senior Member

Following is a response from DHS staff member, Tom Pullum:

Thanks for pointing this out. Yes, it is possible that there was inconsistent treatment of m15=98 ("Other"). I have cross-tabulated m15 with the various m65 variables, which give the reasons for not having a facility ("Institutional") birth. Here's an example (for bidx=1:

. tab m15 m65a,m

| reason didn't deliver
| at health facility:
| cost too much
place of delivery | no yes | Total
----------------------+----------------------+----------
respondent's home | 16,019 2,922 | 18,941
other home | 163 33 | 196
parents' home | 1,730 352 | 2,082
public: govt./munic. | 53,283 0 | 53,283
public: govt. dispens | 3,104 0 | 3,104
public: uhc/uhp/ufwc | 2,416 0 | 2,416
public: chc/rural hos | 41,151 0 | 41,151
public: phc/additiona | 12,621 0 | 12,621
public: sub-centre | 2,072 0 | 2,072
other public sector h | 305 0 | 305
private: hospital/mat | 38,480 0 | 38,480
other private sector | 1,189 0 | 1,189
ngo or trust hospital | 611 0 | 611
other | 372 20 | 392
----------------------+----------------------+----------
Total | 173,516 3,327 | 176,843

In this table, "Other" (m15=98) is classified with home births. You say that in table 8.14 it was classified with institutional births. If so, that would be an inconsistency.

I see an error in the coding of the m65 variables that could have led to this. For example, m65a takes the values 0 and 1 but is never "." for NA. All of the m65 variables should have been coded NA for non-institutional births, corresponding with a skip in the questionnaire. The questions about reasons for not delivering in a facility should only be asked for births that were not in a facility.

I think you have answered your own question. If table 8.14 were calculated correctly, it would not include births for which m15=98. If it does include the births with m15=98, then there is either an error or an inconsistency.

Report message to a moderator

Previous Topic:	Parity progression Ratio
Next Topic:	DHSRates R Standard Errors

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Wed Nov 27 07:59:50 Coordinated Universal Time 2024