Re: st: -prtest- interpreting 1/2 as variable in one sample proportion test

[Jorge Eduardo Pérez Pérez <perez.jorge@ur.edu.co>]

I am aware that putting 0.5 instead of 1/2 works. Still, I believe
-prtest- should be able to interpret the 1/2, as -bitest- does. Both
help files state that the commands take inputs in the #p form, so both
should take the same kind of inputs.

I was able to replicate this in Stata MP 12.1

Thanks !
_______________________
Jorge Eduardo Pérez Pérez


On Wed, Apr 17, 2013 at 3:29 PM, Richard Goldstein
<richgold@ix.netcom.com> wrote:
> for -prtest- you need to .5 rather than 1/2; see the help file
>
> Rich
>
> On 4/17/13 3:20 PM, Jorge Eduardo Pérez Pérez wrote:
>> Dear Statalist.
>>
>> I found the following problem while using -prtest- on Stata S.E. 11.1
>> 64bit , on Windows 7.
>>
>> While running proportion tests, if I run the following binomial test:
>>
>> bys id: bitest above == 1/2  if round<=10
>>
>> The test executes correctly. However, when using the parametric proportion test:
>>
>> bys id: prtest above == 1/2  if round<=10
>>
>> I get:
>>
>> 1/2   is not a 0/1 variable
>> r(450);
>>
>>
>> It seems like -prtest- is misinterpreting the 1/2 and reading it as a
>> string, so it reads as a variable and Stata tries to conduct a
>> two-sample proportion test.
>>
>> This looks like a bug, but I may be missing something. Can someone
>> replicate this in Stata 12? I am working with the 16 feb 2012 version
>> of Stata 11 (The system administrators don't update Stata as they
>> should) Maybe this has been corrected in an update.
>>
>> Thanks,
>> _______________________
>> Jorge Eduardo Pérez Pérez
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/statalist-faq/
> *   http://www.ats.ucla.edu/stat/stata/
>
>


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