Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: error in forval?


From   Austin Nichols <austinnichols@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: error in forval?
Date   Wed, 20 Oct 2010 08:15:30 -0400

helen bian <helenbian06@gmail.com>:

I should have said "range" not "numlist" per the help file for
-forvalues-.  -prtest- is OK only as long as the proportions involved
are far enough away from zero or one and sample sizes sufficient that
confidence intervals boundaries are not near zero or one (note that
the confidence interval on a proportion for a nonzero point estimate
should never include zero).  You might want to check for sensible
results inside your loop by recovering confidence interval bounds for
the individual proportions from -proportion-.

There are cases where prtest, svy:tab, and logit all fail to produce
sensible results; one example is where there is a zero proportion
involved. Perhaps others will suggest their favored approach in such
cases, e.g.

sysuse auto, clear
g y=rep78<3
ren foreign x
prtest y, by(x)
bitest y=`r(P_1)' if x==0
svyset, srs
svy:tab y x, col se
test _b[p21]=_b[p22]
test _b[p21]=0
proportion y, over(x)
mat b=e(b)
mat v=e(V)
di b[1,3]-invttail(e(df_r),.025)*sqrt(v[3,3]),b[1,3]+invttail(e(df_r),.025)*sqrt(v[3,3])
test [_prop_2]_b[Domestic]=[_prop_2]_b[Foreign]
test [_prop_2]_b[Domestic]=0
logit y x, nolog or
binreg y x, or nolog
exlogistic y x, nolog
tab y x
cii 22 0
cii 52 10

On Wed, Oct 20, 2010 at 7:38 AM, Austin Nichols <austinnichols@gmail.com> wrote:
> helen bian <helenbian06@gmail.com>:
>
> Incorrect numlist; try instead
> forvalues i=1(1)6 {
>  forvalues j=`=`i'+1'/7 {
>  di `i', `j'
>  }
> }
>
> but also note that -prtest- is a poor solution for most problems;
> svy:tab is better (even for non-survey data--just -svyset, srs- first)
> or a -logit- of FLAG_ChoseBestHosp1 on the single dummy variable
> defined by BG_Condition_recode==`i'.
>
> On Wed, Oct 20, 2010 at 7:29 AM, helen bian <helenbian06@gmail.com> wrote:
>> Dear all,
>>
>> Could any one point to me what the problem is in the following for loop, please?
>>
>> forvalues i=1(1)6{
>>  forvalues j= `i'+1 /7{
>>  prtest FLAG_ChoseBestHosp1 if ( BG_Condition_recode==`i'|
>> BG_Condition_recode==`j'), by( BG_Condition_recode)
>>     }
>> }
>>
>

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index