The DHS Program User Forum: Dataset use in Stata

Home » Data » Dataset use in Stata » Definition of ARI

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

Definition of ARI [message #13406]

Wed, 01 November 2017 02:51

r10000
Messages: 2
Registered: September 2017
Location: Bangladesh

Member

Dear DHS,

I am trying to analyze the possible risk factor associated with ARI by using BDHS-2014, kr data-set.

DHS uses the following definition of ARI,

gen ari=0
replace ari=1 if h31b==1 & (h31c==1 | h31c==3)
replace ari =. if age>=60 | b5==0

So by this definition of ARI if h31b or h31c has non-eligible cases such as (8 or 9) ,the corresponding observation will be assigned 0 or 'No' in the ari variable.

But the code (8 or 9) doesn't mean that the child did not have symptoms of ARI.

Is this definition of ARI okay to use in regression model?

Report message to a moderator

Re: Definition of ARI [message #13463 is a reply to message #13406]

Mon, 06 November 2017 13:13

Bridgette-DHS
Messages: 3190
Registered: February 2013

Senior Member

Here's a response from Senior DHS Stata Specialist, Tom Pullum:

You are right, cases with h31b or h31c equal to 8 or 9 will be assigned to "ari=0", even though this may be an error. The general rule for most outcomes is that the default category is 0. We only assign to 1 if there is a clear indication. This convention is analogous to favoring the null hypothesis in hypothesis testing, or assuming that someone is innocent unless proven guilty in criminal cases. Favoring "0", if in doubt, may underestimate the prevalence of "1" but is usually considered preferable to doing the reverse.

Report message to a moderator

Re: Definition of ARI [message #13520 is a reply to message #13463]

Mon, 13 November 2017 16:04

r10000
Messages: 2
Registered: September 2017
Location: Bangladesh

Member

Thanks for the reply.

Report message to a moderator

Previous Topic:	Merging and appending Kenya DHS
Next Topic:	Postnatal checkup:Rwanda DHS 2014-15

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Sat Nov 9 09:08:07 Coordinated Universal Time 2024