Looping over observation

[nizam khan]

Hi,
I want to run one-sample test of proportion using following data and save p-val for each obs.
prtesti #obs1 #p1 #p2

I tried the following program. The program is working for fist values only. Can someone tell me where is the problem with my program?
Thanks.

Nizam

gen pval=.
local N = _N
forvalues i = 1/`N' {
prtesti `=obs1' `=p1' `=p2'
local p= 2*(normprob(-abs(r(z))))
replace pval=`p'
}

obs1	p1	p2	pval
1273314	0.5256	0.5150
1631438	0.5278	0.5150
1686524	0.5214	0.5150
1980880	0.5298	0.5120
8187141	0.5158	0.5175
5386442	0.5193	0.5175
4701894	0.4823	0.5173
4232052	0.5311	0.5170

[Nick Cox]

gen pval=.
local N = _N
forvalues i = 1/`N' {
prtesti `=obs1[`i']' `=p1[`i']' `=p2[`i']'
local p= 2*(normprob(-abs(r(z))))
replace pval=`p' in `i'
}

Your prtesti command

1. really doesn't know or care that it is inside a loop.

2. has the same form through the loop. Stata's rule is that the first observation is used in such circumstances when you feed a variable to a command expecting constants.

Note that the replace statement must specify where you want to put the result; otherwise all values are changed..

[nizam khan]

Thanks Nick

[Nick Cox]

See also https://www.stata.com/support/faqs/d...iate-commands/