Re: st: Permute - error message

[Guy Grossman <guygrossman1@gmail.com>]

Dear Nick,
The error message I received is below. However, when running your
adjusted code, the command was executed swiftly.
Thank a million,
Guy

. permute z r(diff): mydiff

(running mydiff on estimation sample)
Warning:  Because mydiff is not an estimation command or does not set e(sample),
          permute has no way to determine which observations are used in
          calculating the statistics and so assumes that all observations are
          used.  This means that no observations will be excluded from the
          resampling because of missing values or other reasons.
          If the assumption is not true, press Break, save the data,
and drop the
          observations that are to be excluded.  Be sure that the dataset in
          memory contains only the relevant data.

Permutation replications (100)
----+--- 1 ---+--- 2 ---+--- 3 ---+--- 4 ---+--- 5
.xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx    50
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx   100
Monte Carlo permutation results                   Number of obs   =          8
      command:  mydiff
        _pm_1:  r(diff)

permute var:  z
------------------------------------------------------------------------------
T            |     T(obs)       c       n   p=c/n   SE(p) [95% Conf. Interval]
-------------+----------------------------------------------------------------
       _pm_1 |   .3586207       1       1  1.0000  0.0000      .025          1
------------------------------------------------------------------------------
Note:  confidence interval is with respect to p=c/n.
Note:  c = #{|T| >= |T(obs)|}
Note:  missing values observed in permutation replicates.

On Thu, Mar 31, 2011 at 8:28 AM, Nick Cox <njcoxstata@gmail.com> wrote:
>
> I don't know what error messages you are getting. You don't tell us
> and what I did, as below, ran without error.
>
> The first thing I noticed was that your code can be slimmed down. It
> seems to me that
>
> > egen grouptotal = rowtotal(y1-y10)
> > egen groupN = rownonmiss(y1-y10)
> >
> > cap program drop mydiff
> > program define mydiff, rclass
> >        egen tot_treat = total(grouptotal) if z ==1
> >        egen tot_cont = total(grouptotal) if z ==0
> >        egen N_treat = total(groupN) if z ==1
> >        egen N_cont = total(groupN) if z ==0
> >        sum tot_treat
> >        local size_treat=r(max)
> >        sum tot_cont
> >        local size_cont=r(max)
> >        sum N_treat
> >        local N_treatment=r(max)
> >        sum N_cont
> >        local N_control=r(max)
> >        return scalar diff =
> > (`size_treat'/`N_treatment')-(`size_cont'/`N_control')
> > end
>
> is equivalent to
>
> egen mean = rowmean(y1-y10)
> cap program drop mydiff
> program mydiff, rclass
>       su mean if z == 1, meanonly
>       local mean_treat = r(mean)
>       su mean if z == 0, meanonly
>       return scalar diff = `mean_treat' -  r(mean)
> end
>
> I am a big fan of -egen- when appropriate, but egenitis is a condition
> in which it is used quite unnecessarily to create variables merely to
> hold single constants.
>
> When I ran that I got output with -permute-.
>
> That said, I don't think that I what I did was anything other than
> making style changes, so I can't explain your report of errors.
>
> Nick
>
> On Thu, Mar 31, 2011 at 11:40 AM, Guy Grossman <guygrossman1@gmail.com> wrote:
>
> > I am trying to conduct a simple randomization inference analysis
> > (permuting Z - treatment indicator). y1-y10 are outcomes of individual
> > units within groups identified by the id variable. The number of units
> > within groups varies from 3 to 10. In the original data 3 groups are
> > treated (z=1) and 5 are control (z=0).
> >
> > For some reason whether I am using the permute command or the
> > user-written "tsrtest" command (by Kaiser and Lacy) I get error
> > messages which I fail to correct. Any suggestions on how to get this
> > code running would be highly appreciated.
> >
> > Thanks!
> > Guy
> >
> > input id z y1 y2 y3 y4 y5 y6 y7 y8 y9 y10
> > 1 0 0 0 0 0 0 . . . . .
> > 2 0 0 0 0 0 0 0 0 0 0 0
> > 3 0 1 1 1 1 0 . . . . .
> > 4 0 0 1 1 . . . . . . .
> > 5 1 0 0 1 1 . . . . . .
> > 6 1 1 1 1 1 1 . . . . .
> > 7 0 0 0 0 1 0 0 . . . .
> > 8 1 0 0 0 1 0 1 . . . .
> > end
> >
> > egen grouptotal = rowtotal(y1-y10)
> > egen groupN = rownonmiss(y1-y10)
> >
> > cap program drop mydiff
> > program define mydiff, rclass
> >        egen tot_treat = total(grouptotal) if z ==1
> >        egen tot_cont = total(grouptotal) if z ==0
> >        egen N_treat = total(groupN) if z ==1
> >        egen N_cont = total(groupN) if z ==0
> >        sum tot_treat
> >        local size_treat=r(max)
> >        sum tot_cont
> >        local size_cont=r(max)
> >        sum N_treat
> >        local N_treatment=r(max)
> >        sum N_cont
> >        local N_control=r(max)
> >        return scalar diff =
> > (`size_treat'/`N_treatment')-(`size_cont'/`N_control')
> > end
> > tsrtest z r(diff): mydiff
> > permute z r(diff): mydiff
>
> *
> *   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/

*
*   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/