Re: Injuries [message #24325 is a reply to message #24317] |
Fri, 22 April 2022 09:10 ![Go to previous message Go to previous message](/theme/default/images/up.png) |
Janet-DHS
Messages: 802 Registered: April 2022
|
Senior Member |
|
|
Following is response from DHS Research & Data Analysis Director, Tom Pullum:
The number 87,929 is the weighted number of de facto residents. Open the PR file with Stata and enter this line "tab hv102 hv103 [iweight=hv005/1000000]". You will get the following table:
. tab hv102 hv103 [iweight=hv005/1000000]
usual | slept last night
resident | no yes | Total
-----------+----------------------+----------
no | 0 1,994.866 | 1,994.866
yes | 2,884.21 85,934.39 | 88,818.6
-----------+----------------------+----------
Total | 2,884.21 87,929.26 | 90,813.47
You see 87,929 as the (weighted and rounded) total in the second column. This is like a population total for the denominator of a death rate. All of the numbers in the last column of tables 2.18 and 2.21 come from the PR file and serve as the denominators of the rates in the rows of those tables.
However, none of these people died before the survey, because of traffic accidents or anything else. (Some people may have been injured.) To get at deaths or injuries you have to go to the HR file. In general if something was included in the survey and you simply can't find it anywhere else, you can check the HR file, which has one record per household, and you may find it there.
Open the HR file and enter the following lines in Stata:
keep hv001 hv002 hv005 sa*_*
reshape long saidxa_ sa04_ sa05_ sa06_ sa06a_ sa07_ sa08_ sa09_ sa10a_ sa10b_ sa10c_ sa10d_ sa10e_ sa10f_ sa10g_ sa10h_ sa10i_ sa10j_ sa10k_ sa10l_ sa10x_ saidxi_ sa15_ sa16_ sa17_ sa17a_ sa18_ sa19_ sa20_ sa21a_ sa21b_ sa21c_ sa21d_ sa21e_ sa21f_ sa21g_ sa21h_ sa21i_ sa21j_ sa21k_ sa21l_ sa21x_ sa26_ sa27_ sa28_ sa29_ sa30_ , i(hv001 hv002) j(case)
rename *_ *
tab saidxa,m
drop if saidxa==.
In the reshape step the labels will be lost and you need to re-attach them, but otherwise this gives you a file of the deaths and injuries. You can use it to construct the numerators of the rates (be sure to weight with hv005). The rates are constructed as 100,000*numerator/denominator.
There are more elegant ways to get this file but this way should work. Let us know if you have other questions.
|
|
|